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
More warning bits that never got committed.
More appropriate compiler flags for warning checks (macosx only for the moment).
The changes in dgn*[lc] just rename line_number to nh_line_number to avoid a
clash, so no need to regenerate the lex output.
get macosx down to one hints file (default tty, single user) for
tty, x11, qt and single or multiple users
preliminary bits that might allow a macosx qt build
Add MAXPLAYERS to SYSCF config file; deprecate (but continue to support)
MAX_NR_OF_PLAYERS in nethack.sh since it is trivially overridden in many
(all?) cases and isn't useful for ports that don't use nethack.sh.
infrastructure for "system options" - things currently specified at build
time that should be changeable at install time or run time but not really
under user control
generalize contact info so it can be localized and it doesn't have to be
an email address
move recently introduced WIZARDS into sysopt
drop bogus OPTIONS=wizards possibility
new function build_english_list() to comma-ize and add 'or' from a whitespace separated list: A. A or B. A, B, or C.
syscf file now handles: WIZARDS SUPPORT RECOVER
SUPPORT specifies local support information
RECOVER will eventually supply port-specific and/or localized info on how
to run recover (or get it run for you).
Note: in sys/msdos I changed sys.o (generated from pcsys.c) to pcsys.o
Note: sys/msdos/Makefile.GCC has 2 rules for sys.o (now pcsys.o)
Add options SYSCF (to add a system-wide configuration file) and SYSCF_FILE
(for a file-based implementation of SYSCF) - this allows a binary distribution
to be configured at install time. The only option supported at this time is
WIZARDS - a list of usernames which can use -D; currently only for unix-likes
but should be extendable to anything that has a concept of multiple users.
Add mac tty multiuser config using sgid.
From a bug report.
Pointer subtraction yields a result of type ptrdiff_t which can be
bigger than type long, potentially resulting in warnings about value
truncation--in these cases, when assigning to a long.
There was an issue reported where save files between different
versions of a manufacturer's compiler were incompatible because the time_t
ubirthday field was changed from 32 bits to 64 bits.
32 bit time_t implementations will break at 19:14:07 on January 18, 2038.
64 bit time_t implementations will break at 23:59:59 on December 31, 3000.
This removes the dependency on the size of time_t from the save file.
The ubirthday field is no longer embedded in struct you.
This also adds two general purpose routines to hacklib.c, one to convert a time
value to a 14 character char representation and the other to convert that
back to time_t. Those are used by the save/restore routines.
This is a savefile breaking change, so editlevel in patchlevel.h was
incremented.
Add a hints file for an ubuntu tty build and skip xset if no fonts.dir
file exists (this test will probably need another tweak, but I need a
linux x11 build first to test it).
build system.
Anyone who wants to do a build from sys/unix and doesn't want to figure this
out just needs to do:
sh setup.sh hints/unix
instead of:
sh setup.sh
and then continue on as usual.
New files:
sys/unix/NewInstall.unx - the new directions
sys/unix/hints/* - the hints files. There will be more later.
sys/unix/mkmkfile.sh - helper for setup.sh
Summary of changes:
see NewInstall.unx for info on the new build system
introduction of various preprocessor symbols to turn options off that
are defaulted on historically
comment out nethackrc (and similar) entries that still use the old symbol
syntax.
commenting out of Makefile.* lines that now come from hints/unix
GAMEDIR is replaced with HACKDIR so the Makefiles and the C source agree.
Note that I have NOT changed the docs and/or Makefiles for be, msdos, os2,
vms, or winnt. If port maintainers don't then I will, but I can't test
those ports.
nethack.sh now handles the font path automatically
The fix for #H1409 a couple of days ago included updates for
sys/vms/Makefile.* but not vmsbuild.com. I thought the latter didn't
need any, but was mistaken; vmsfiles.c needs to be compiled early in
order for vmsfiles.obj to be present in the object library when makedefs
gets linked.
Fix the problem From a bug report. His system has a logical name "DATA" pointing at some disk, and
when the dlb utility tried to open "data" for inclusion in the library
being built at install time, it attempted to access the wrong thing and
failed. He then attempted to fix it in a manner which let dlb finish, by
modifying dlb_main.c to append "." to file names that lack a dot, but
then nethack couldn't access "dungeon" in the library because string
comparison didn't match the altered dlb directory entry of "dungeon.".
NetHack was working around this unintended interaction with the
environment issue in fopen_datafile(), and dlb was doing so for fopen()
but not open(). This moves nethack's fixup out of src/files.c and into
sys/vms/vmsfiles.c, adds another routine there so that both open() and
fopen() are covered, and updates the vms Makefiles so that the various
utility programs all link with vmsfiles. (The build script vmsbuild.com
puts object files into a library and gets that last bit for free.)
Last fall when Michael added the symset stuff to supersede the old
handling for IBMgraphics and DECgraphics, Guidebook.tex was changed to
support multi-page tables in the output. But that requires that the
input be processed twice, because it requires feedback stored in
Guidebook.aux and the first pass can't rely on that file being present
or up to date. This updates the Unix, VMS, and OS/2 makefiles to do
two-pass processing. (I didn't see latex usage anywhere else, and the
branch version doesn't include the formatting change which needs this.)
Suggested by <Someone>: in the special level compiler, support
"random" in addition to "open" and "closed" for a drawbridge's inital
state. Drawbridge shares code with door, so the necessary parsing was
already present. This just treats random as valid for drawbridge instead
of explicitly rejecting it, and makes the special level loader process it.
He also suggested that the two drawbridges on the bottom level of
the Valk's quest be changed from open to random, but this patch doesn't
go that far. I think it's a good idea, but since the player can't use a
musicial instrument on those bridges, this has more impact on game play
than it might at first seem. I don't really want to see Valkyries be
required to use magic for occasions where both bridges start out closed.
Reorganize the recent wizard mode control: move set_playmode() from
xxxmain.c to the core, and have it call new authorize_wizard_mode() to do
the port-specific part. If the set_playmode() call during startup doesn't
result in running in wizard mode (either because not allowed or user
didn't request it), it will be called again during restore if the save
file is from a wizard mode game.
For ports which check character name for authorization, players will
have to use `nethack -u whatever -D' (or options for name and playmode) to
restore a wizard mode save file if WIZARD has been changed from "wizard".
plname[] from a wizard mode saved game will always have that value, so if
it's not the right one players will need to get authorized by the startup
code before loading the save file.
Wizard mode or explore mode can be forced on (via -D or -X on the
command line, or now via OPTIONS=playmode:debug|explore) when restoring
a saved game; explore mode handling was confined to restgamestate(), but
wizard mode handling was replicated in every main(). Treat `wizard' the
same as `discover'. Also, prevent a new game started when restore fails
from using the old game's option settings if partial restore attempt got
far enough to load the flags struct. And update bemain.c and macmain.c
to catch up with the others modified by the playmode patch.
[see cvs log for src/options.c for some additional info]
Relief for the command-line impaired. Allow player to request
explore or wizard mode via run-time config file or NETHACKOPTIONS.
Validation is left to xxxmain() and has been updated for Unix, VMS, and
ports which share pcmain. Mac and Be appear to allow any user to access
wizard mode, and may not need any modification, although they'll continue
to have the old buglet of running with both wizard and discover flags set
if player uses `nethack -X -D'. This may or may not work as-is for the
Qt interface depending upon whether it goes through one of the xxxmain()'s
mentioned above [someone needs to make sure that it doesn't allow Qt on
Unix to bypass the (username == WIZARD_NAME) test when user requests
wizard mode].
NOSAVEONHANGUP isn't documented anywhere. I don't see why it should
wipe out recoverable level files just because it doesn't want to build a
save file out of them during the hangup. Leave them intact if checkpoint
is active. If someone really wants to make them always go away, they'll
need to disable INSURANCE as well as enable NOSAVEONHANGUP.
tty's getret() -> xwaitforspace() could get stuck in a loop after
hangup, depending upon the state of terminal shutdown (accepting EOF or
ESC cares about cbreak mode?). Make xwaitforspace() become a no-op during
hangup processing.
vms's call to hangup() from an exit handler took place after the
terminal has been reset (the exit handler for the latter is registered
later so executes sooner). Then exit_nhwindows() resulted in a second tty
reset attempt and settty() -> setctty() encountered an error (which it
reported, triggeting a getret() call). Make the vms code correctly guard
against multiple resets.
[See cvs log for src/cmd.c for more complete description.]
This turns clearlocks() into a no-op during the period when the UNIX
port is asking the user to confirm whether to overwrite an existing game.
Also, this removes the duplication of code and function between hangup()
and end_of_input(), and it simplifies the check for whether hangups are
supported by adding new macro HANGUPHANDLING. (I don't think global.h is
the best place to be defining that but I couldn't figure out where else
it would fit, other than repeating for individual xxxconf.h files.) And
adds a couple more done_hup checks to try to cope with situations where
rhack() is being bypassed. Lastly, having readchar() return EOF was
ignored for non-UNIX configs; now everybody gets ESC instead of letting
EOF be seen further inside the core.
- force time_t to be 32 bits via Makefile. (If struct u gets changed to not
store a time_t field, this can be removed).
- Latest Windows SDK (supporting Vista) causes a warning due to
a prototype in sys/winnt/nttty. Since we don't actually link with that
function and rely on a DLL find procedure at run time, comment out
the prototype. The function is now officially in the SDK in the
windows header files anyway.
Explicitly truncate the query prompt string to QBUFSZ-1 characters.
For tty and Amiga, no longer include the choices and default within that
length limit; use a bigger buffer to hold them along with the prompt.
[See cvs log for doc/window.doc for more details.]
This patch attempts to add some levels of unicode support
to NetHack.
The master on/off switch for any Unicode support is
defining UNICODE_SUPPORT in config.h. Currently
there is code support for two subsets of unicode support:
UNICODE_DRAWING
If UNICODE_DRAWING is defined, then the data
structures used to house drawing symbols are expanded
to the size of wchar_t, big enough to hold unicode characters.
A typdef called `nhsym' is involved and if UNICODE_DRAWING
is defined, it is wchar_t, otherwise it is uchar.
UNICODE_WIDEWINPORT
If UNICODE_WIDEWINPORT is defined, then the data
structures inside the window port are expanded to the size of
wchar_t, big enough to hold unicode characters. Both map
symbols and text within the window port are expanded, in order
for potential support for displaying multinational characters some
day, but this patch only provides viewing of map symbols.
A typdef called `nhwchar' is involved and if UNICODE_WIDEWINPORT
is defined, it is wchar_t, otherwise it is char.
The only window port with code support for UNICODE_WIDEWINPORT
currently is the TTY port. Don't enable UNICODE_WIDEWINPORT
unless:
- it is a TTY port
- the underlying platform specific routines can
handle the larger data structures.
Don't enable UNICODE_SUPPORT unless:
- your compiler can handle wchar_t.
- your compiler can accept L'a' characters.
- your compiler can accept L"wide" strings.
Note that if your compiler can handle the above, you could
enable the larger data structures (currently if TTY) even if your
platform can't actually display unicode or UTF-8, by messing
with u_putch() in win/tty/wintty.c to only deal regular chars.
That should be the only function that actually pushes wide characters
out to the display.
If you enable UNICODE_SUPPORT, and your platform is capable
you will need to turn on the unicode run-time option to be able to
load unicode character sets from the symbol file, to be able to
push unicode characters to the display. You'll also want to load
a unicode symbol set once the unicode option is toggled on. In
a config file you would do that via these two lines:
OPTIONS=unicode
OPTIONS=symset:Unicode_non_US
The repository was stamped with NETHACK_PRE_UNICODE
prior to applying this patch, and stamped with
NETHACK_POST_UNICODE afterwards. The code differences
between those two tagged versions are this patch.
Add putmixed() to the window port. It allows map symbols to
be included in the string by encoding them in a unique fashion.
This was done because Unicode symbols, for instance, could be
longer than the size of a char.
The encoding of the map symbols in this patch is done by
prefixing a glyph value with \GXXXX, where XXXX is a
random value for the current game. The reason for the random
prefix is to minimize the possibility that a player can trigger
the escape sequence processing within text under their control
(dog names, etc.) the way they could if the sequence was fixed
in the source code. The random prefix remains the same throughout
the lifetime of a game because message window strings are
saved in the save file.
(There was actually a bug present because of the embedded
character even before the recent symbol changes, because if
someone was using a different set of characters between games,
the saved messages would reflect the original characters, rather
than the current. That bug was introduced with the ability to
save messages to the savefile.)
A window port does not have to supply an XXX_putmixed() routine,
it can use genl_putmixed() which uses the old behavior of
embedding the sequence as a character within the string
and calling putstr(). genl_putmixed() takes care of the decoding
of the escape sequence.
This also #ifdef's out code in pager.c for converting a glyph
to a character, and uses mapglyph() to do that instead. Does
anyone see a problem with doing that through mapglyph instead
of repeating similar code within pager.c?
The palette option is supposed to be allowed in the config file
without a value for win32 to trigger a load of a predefined
NetHack palette, but that wasn't working.
This fixes that. To prevent the use of any palette modification
code at all, just leave the palette option out of the config
file entirely.
Clean up the preprocessing associated with the
loadable symbol stuff.
Base it on new LOADSYMSETS, rather than on the
previously existing ASCIIGRAPH preprocessor define.
- reduce the number of symbol tables for each graphics
set {PRIMARY, ROGUESET} from three {map, oc, mon}
tables for each of the display symbols, the loadable symbols,
and the rogue symbols, to one continguous table for
each:
showsyms: the current display symbols
l_syms: the loaded, alterable symbols
r_syms: the rogue symbols
- Modify mapglyph so that the index into the symbolt table is
available as a return value (it was a void function), rather than
just the char converted from the glyph.
- That makes it possible for a window port to use the same
index value to extract from another table (perhaps a unicode
table) for a different set of display symbols. The index
is much more useful than trying to convert the character
into another type of symbol, as some contributed patches
have done.
- It is much easier to load a single alternative flat table to
make substitutions, since the corresponding value just
has to get placed into the same index offset in the
alternative table.
This also fixes a bug I found in botl.c, where you could
go to the rogue level, and the bottom line gold symbol
was not being updated with the new character as it should.
The reason was because the gold value had not changed,
only the field symbol used had changed.
This updates multiple ports to place a (void) cast on
the mapglyph call, now that it returns a value, so this
is going to generate a lot of diff e-mails.
Pat Rankin wrote:
> Symbol set definitions need a description attribute, above and
> beyond allowing comments in the file, for inclusion in the 'O'
> command's menu entries for selecting them.
[...]
> mapglyph.c isn't the proper place to decide whether to define
> ROGUE_COLOR. That may need to become a symbol attribute,
> which we'd then specify on the Epyx rogue set(s).
Implement both of the suggestions above.
This takes the PC config file commented symbol value
recommendations from <Someone> for blind players
and puts them into a symset.
[note to devteam: They look odd. I thought perhaps that
something was code wrong, but I went back to 3.4.3
and uncommented the config file stuff. They look the
same there, still odd, especially corridors.
Does anyone have any of the e-mail from <Someone> that might give an indication of what is supposed
to be seen on the display? I wonder if those config
file options fell out of synch with the code long ago]
- Instead of checking for the Rogue level, check which
graphics are engaged (PRIMARY or ROGUESET) in the
SYMHANDLING() macro.
- track which graphics are active through 'currentgraphics'.
- Instead of symset and roguesymset and symhandling and roguehandling
variables, have symset and symhandling be arrays of two, with the
following indexes:
PRIMARY
ROGUESET
That reduced the amount of repeated code.
(Not to be confused with the 'symset' and 'roguesymset' config file options
both of which still exist)
- the symbol routines were adjusted to pass
the index , rather than 'rogueflag' and coding to roguesymset etc.
Other than fixing bugs that are encountered, this is probably
the last of the symbol stuff, with the exception of
making the symset and roguesymset config file options
accept the keyword value "default".
- tile2x11 would not build because drawing.c now depended on strcmpi which
was (via STRNCMPI not being defined) defined to strncmpi which is
implemented in hacklib.c which needs panic which is defined in... I gave up
on tracking down all the loose ends and changed the strcmpi to strcmp,
which means the handling is case sensitive, but it avoids a bunch of
changes to the way the util/Makefile.
- the symhandling changes introduced a chicken and the egg problem for
ASCIIGRAPH on Unix platforms, which was getting the defn from tcap.h but
that does not get included earlier enough nor often enough. I added a defn
to unixconf.h to mimic ntconf.h, since ASCIIGRAPH is normally defined on Unix.
- options.c included an unused decl for a function named graphics_opts
- Unix Makefile was not installing "symbols". I'm assuming this isn't
supposed to get the DLB treatment.
