Qt menus have [ok][cancel][all][none][other stuff] buttons across
the top but it was disabling [cancel] for inventory viewing and
other pick-none menus. Enable that so that [cancel] is a viable
alternative to typing ESC or clicking on [ok] for dismissing the
menu without picking anything.
This greatly improves the '?' choice when using the '#' prefix
to select an extended command. It isn't perfect, because the
<text><spaces><more text> layout (using spaces to pad the first
column so that the second one lines up) produces at least one
line where <more text> is so long that it wraps, and instead of
| <text><spaces><start of more text>
| <blank><spaces><rest of more text>
that selectable menu entries have, the continuation is
| <text><spaces><start of more text>
|<rest of more text>
(made slightly worse by the fact that <text> is indented a little
and <rest...> isn't.
This affects the aesthetics of all NHW_MENU windows, not just the
one that desparately needed help. Maybe the core should send some
hint on a menu or text window by menu or text window basis about
whether or not fixed-width font is preferable to variable one.
Qt's implementation of '#' puts up a rectangular grid of buttons
containing command names from the alphabetized extcmdlist[]:
| # ? adjust annotate
| apply attributes autopickup call
| cast ...
When 3.6 put all commands into that list, the hardcoded 4 columns
resulted in so many rows that the grid wouldn't fit on the screen
(at least not on my smallish laptop screen). There's no scrollbar
so the commands beyond "takeoff" were inaccessible off the bottom.
Warning messages from within Qt were issued to stderr complaining
about trying to render something off the screen (once each time the
'#' command grid was generated).
It was also including wizard mode commands when not in wizard mode.
Suppress those when they're not applicable, and change the grid to
use 6 columns then and 8 for wizard mode. The appropriate amount
ought to be calculated on the fly but these values work ok with the
current command list. (On my screen; if something smaller is used,
the original problem could come back, just not as severe as before.)
Having an alphabetized list go across rows instead of down columns
feels counter-intuitive so transpose the grid.
| # autopickup ...
| ? call
| adjust cast
| annotate ...
| apply
[Having another button next to <cancel> that lets the user switch
back and forth between the two orientations could be worthwhile.
A full-fledged wc/wc2 option for that doesn't seem warranted.]
The commands can be selected by typing their names as an alternative
to mouse click. The input widget supports <backspace> but lacked
handling for <delete> so add that.
When typing a command by its name, a new grid showing only matching
candidates gets displayed so that you can switch back to mouse input.
It looks pretty bad but does work as intended. I didn't touch that;
however, it looks different now due to the columns-vs-rows change.
The menu after picking "?" looks worse. It assumes a fixed width
font and tries to align things in two columns with spaces, but the
result when using a variable width font is ugly. This makes no
attempt to address that.
For Qt, if unable to load either nhfiles.bmp or x11tiles, quit after
giving the can't-load-tiles feedback instead of continuing on and
eventually triggering a segfault.
In file included from ../win/Qt/qt_bind.cpp:20:
In file included from /usr/local/Cellar/qt/5.15.0/lib/QtGui.framework/Headers/QtGui:3:
In file included from /usr/local/Cellar/qt/5.15.0/lib/QtGui.framework/Headers/QtGuiDepends:3:
In file included from /usr/local/Cellar/qt/5.15.0/lib/QtCore.framework/Headers/QtCore:4:
In file included from /usr/local/Cellar/qt/5.15.0/lib/QtCore.framework/Headers/qglobal.h:1302:
/usr/local/Cellar/qt/5.15.0/lib/QtCore.framework/Headers/qflags.h:121:41: warning: declaration shadows a
variable in the global namespace [-Wshadow]
Q_DECL_CONSTEXPR inline QFlags(Enum flags) noexcept : i(Int(flags)) {}
^
[…]
../include/flag.h:390:29: note: previous declaration is here
extern NEARDATA struct flag flags;
^
In file included from ../win/Qt/qt_click.cpp:18:
In file included from /usr/local/Cellar/qt/5.15.0/lib/QtGui.framework/Headers/QtGui:3:
In file included from /usr/local/Cellar/qt/5.15.0/lib/QtGui.framework/Headers/QtGuiDepends:3:
In file included from /usr/local/Cellar/qt/5.15.0/lib/QtCore.framework/Headers/QtCore:36:
/usr/local/Cellar/qt/5.15.0/lib/QtCore.framework/Headers/qcache.h:191:15: warning: declaration shadows a
variable in the global namespace [-Wshadow]
Node *u = n;
^
../include/decl.h:219:23: note: previous declaration is here
E NEARDATA struct you u;
^
[…]
In file included from ../win/Qt/qt_click.cpp:18:
In file included from /usr/local/Cellar/qt/5.15.0/lib/QtGui.framework/Headers/QtGui:5:
In file included from /usr/local/Cellar/qt/5.15.0/lib/QtGui.framework/Headers/qabstracttextdocumentlayout.h:45:
In file included from /usr/local/Cellar/qt/5.15.0/lib/QtGui.framework/Headers/qtextlayout.h:47:
In file included from /usr/local/Cellar/qt/5.15.0/lib/QtGui.framework/Headers/qcolor.h:44:
/usr/local/Cellar/qt/5.15.0/lib/QtGui.framework/Headers/qrgb.h:66:46: warning: declaration shadows a
variable in the global namespace [-Wshadow]
inline Q_DECL_CONSTEXPR QRgb qRgb(int r, int g, int b)// set RGB value
^
../include/decl.h:1208:27: note: previous declaration is here
E struct instance_globals g;
^
[…]
In file included from ../win/Qt/qt_glyph.cpp:21:
In file included from /usr/local/Cellar/qt/5.15.0/lib/QtGui.framework/Headers/QtGui:5:
In file included from /usr/local/Cellar/qt/5.15.0/lib/QtGui.framework/Headers/qabstracttextdocumentlayout.h:48:
/usr/local/Cellar/qt/5.15.0/lib/QtGui.framework/Headers/qpalette.h:107:49: warning: declaration shadows a
variable in the global namespace [-Wshadow]
inline void setCurrentColorGroup(ColorGroup cg) { data.current_group = cg; }
^
../include/decl.h:1216:30: note: previous declaration is here
E const struct const_globals cg;
^
fixes#361
Also, experminental introduction of vt_sounddata to enable tty to pass
a sound file index to the terminal side of things where perhaps someone
can add code to something like hterm to take the information relayed by
NetHack to trigger user_sounds locally even if playing on a server.
Compile time option TTY_SOUND_ESCCODES required to build that support in.
It should be independent of TTY_TILE_ESCCODES.
Provide a way to communicate additional behaviors and/or appearances
desired from NetHack window port menus.
This is foundation work for changes to follow at a future date.
groundwork only - window port interface change
This changes the last parameter for add_menu() from a boolean
to an unsigned int, to allow additional itemflags in future
beyond just the "preselected" that the original boolean offered.
There shouldn't be any functionality changes with this groundwork-only
change, and if there are it is unintentional and should be reported.
In order for 'make depend' to be able to handle both Qt4/5 and Qt3,
they need to operate on different object file names.
renames qt*.o to qt3*.o for Qt3
renames qt*.cpp to qt3*.cpp for Qt3 (not essential but seems worthwhile)
moves Qt3's headers from include/qt*.h to win/Qt3/qt3*.h
copies include/qt_xpms.h (before rename) or win/Qt3/qt3_xpms.h (after)
to win/Qt/qt_xpms.h so that Qt4/5 no longer shares one header file
modifies win/Qt3/*.cpp and win/Qt3/qt3_win.h to reflect new header names
modifies Makefile.src to have Qt3 'moc' commands use new names
updates Makefile.src via re-running 'make depend'
'make depend' was only looking at include/*.h to find nested inclusion.
Now it will also look at win/*/*.h. That found a bunch of missing
dependencies for the old gnome sources and a few for Qt3.
Building without Qt still works. Building with it (any version) has
not been tested.
Performance profiling showed that multiple strcmpi() calls were
occurring each and every time a character was going to the map.
This update:
- honors the WC_COLOR capability
- It allows a window-port to control individual color availability should the window-port wish to do so.
- Makes checking on the individual colors for the active window-port is a straightforward table lookup at the CLR_ offset.
iflags.use_color remains a master on/off switch for use of color, regardless of the capability
compiled into the game (default TRUE).
The has_color() routine, which is now a shared routine in src/windows.c, could likely be made
into a simple macro to eliminate the function call, but this update does not go that far.
This hits a lot of port files due to the window-port interface change, mostly cookie-cutter.
Quiting without ever examining inventory caused the Qt interface
to issue an impossible(), then crash due to deferencing a Null
pointer. The prior fix was to suppress the validation code that
was crashing. This changes things so that the inventory window
always gets at least one use, allowing the Qt validation code to
succeed. tty and X11 are ok with it; win32 needs to be verified.
Put in Ron Vaniwaarden's fix for crash occuring if player quits
right after choosing a character and declines to disclose anything.
Just a guess, but deleting the never-viewed inventory window might
be the cause. (3.4.3 didn't delete it.)
There's a small amount of reformatting, but cleaning that up
manually is a monumental task.
Changes to be committed:
modified: doc/window.doc
modified: include/qt_win.h
modified: include/trampoli.h
modified: include/winX.h
modified: include/wingem.h
modified: include/winprocs.h
modified: include/wintty.h
modified: src/display.c
modified: src/windows.c
modified: sys/amiga/winami.p
modified: sys/amiga/winfuncs.c
modified: sys/amiga/winproto.h
modified: sys/wince/mswproc.c
modified: sys/wince/winMS.h
modified: win/Qt/qt_win.cpp
modified: win/X11/winmap.c
modified: win/chain/wc_chainin.c
modified: win/chain/wc_chainout.c
modified: win/chain/wc_trace.c
modified: win/gem/wingem.c
modified: win/gem/wingem1.c
modified: win/gnome/gnbind.c
modified: win/tty/wintty.c
modified: win/win32/mswproc.c
modified: win/win32/winMS.h
print_glyph now takes a second parameter.
Tiles on tiled ports always looked odd on places like the plane of air
where the background color of the tile didn't match the general background
of the surrounding area.
3.6 made that even worse and more glaringly noticeable with the introduction
of darkened room tiles.
The code to actually send something useful through the new parameter
for window ports to take advantage if they want will follow.
I did my best to exempt some of the bigger aligned blocks from the reformatting
using the /* clang-format off */ and /* clang-format on */ tags. Probably some
that shouldn't have been formatted were anyway; if you encounter them, please
fix.
The clang-format tags were left in on the basis that it's much easier to prune
those out later than to put them back in, and it means that, modulo my custom
version of clang-format, I should be able to run clang-format on the source tree
again without changing anything, now that Pat has fixed the VA_DECL issues.
There is a lot of code affected by this, and Pat Rankin correctly
observes that it would be better to store roguelike as a level flag
rather than just using Is_rogue_level. A note for the future.
Part II of the bones tracking patch. Change umpteen different outrip()
routines to handle its new time_t argument, and use formatkiller() instead
of directly accessing killer.{format,name} and killed_by_prefix[]. The
latter is now static within formatkiller().
The many sys/* and win/* changes are untested....
