代码拉取完成，页面将自动刷新
SDL2 variant: dimmed, bold, blinking, and RGB text are handled, as is window resizing.
DOS variant: cross-compilation from GNU/Linux works with both DJGPP and Watcom C/C++; and, 16-bit and 32-bit Watcom makefiles have been combined into one
Many modifications (and some bug fixes) taken from William McBrine's branch. An effort has been made to narrow the gap between the forks where possible. But both forks are moving targets with different design choices and priorities.
Demos improved to show version info
compilation warnings/errors with some compilers in some variants #53, #57, #58, #65, ...
newtest sample was broken in all widw variants #60
the paste button printed debug output #62
some corner cases (midnight crossing, atomicity of tick count reads) in DOS version of napms() were not handled well
New WinGUI (Windows GUI) and SDL2 backends. SDL1 is still supported, but may eventually go away.
Bold, italic, underlined, overlined, dimmed, 'strikeout', blinking text, 256 colors and color pairs, and full RGB colors. These are all supported in WinGUI and mostly supported in X11, SDL1 and SDL2.
In WinGUI, one can choose a font, and both programmatic and user resizing are supported. (Recompiling is necessary to change the font in X11.)
(WinGUI only) Support of SMP Unicode (points beyond 64K) and combining characters. This may be extended to X11 and SDL2 eventually.
Demos corrected to be buildable and testable with ncurses.
(note that not all of these are available on all backends)
Support for up to nine mouse buttons and wheel and tilt-wheel mice, and double and triple mouse clicks
(X11, WinGUI, Win32) Extended range of keys that are recognized. This now includes most of the "oddball" keys such as 'browser back' and 'favorites' found on some keyboards.
Blinking cursors in WinGUI and X11 of various shapes (this could be extended to SDLx eventually).
In X11 and WinGUI, one can call resize_term( ) before initscr( ) to set the initial window size.
Soft Label Keys (SLKs) are considerably more flexible, with the ability to set arbitrary numbers of keys and groupings. See slk.c for details. This applies to all backends.
Many changes to testcurs to test all these new features, and newtest
added to test still more features.
Option to get detailed version information of the used PDCurses library
at run time with new exported PDC_version as PDC_version_info structure.
ACS_x and WACS_x #defines extended to include a lot of "alternative
characters" that overlap in Unicode and CP-437: double-line box chars,
card suits, vulgar fractions, etc.
This applies to all backends. See acs_defs.h for the full list.
Cleaned up some makefiles for Win32 and WinGUI.
On both platforms, vcwin32.mak can now be used with the Intel(R) compiler,
and mingwin32.mak can be used to cross-compile from Linux, or in
command.com under Windows, or with Cygwin/MSYS.
Also added a makefile for Digital Mars for the DOS version.
The "def" files that were needed before to create PDCurses on Windows are removed as they are no longer necessary.
Nothing much new this time, but I've been sitting on some bug fixes for almost a year, so it's overdue. Apart from bugs, the main changes are in the documentation.
In x11, the xc_atrtab table size was under-calculated by half, resulting in crashes at (oddly) certain line counts. (It should've crashed a lot more.) Reported by Mark Hessling.
Test for moved cursor was omitting the window origin offset. Reported by Carey Evans.
Is DOS and OS/2, the value for max items in key_table was still wrong. Reported by C.E.
Changed isendwin() so it won't crash after delscreen().
Ensure zero-termination in PDC_mbstowcs() and PDC_wcstombs().
Disable QuickEdit Mode when enabling mouse input for the Win32 console; reported by "Zalapkrakna".
Fix for building under Innotek C (I hope). Report by Elbert Pol, fix courtesy of Paul Smedley.
Unified exports list with no duplicates -- pdcurses.def is now built from components at compile time.
Don't install curspriv.h, and don't include it with binary distributions.
Building DLLs with LCC is no longer supported, due to the primitive nature of its make.exe.
Export the terminfo stub functions from the DLLs, too.
Added support for Apple's ".dylib" in configure. Suggested by Marc Vaillant (who says it's needed with OS 10.5.)
In sdl1/Makefile.mng, ensure that CC is set.
In the gcc makefiles, "$?" didn't really have the desired effect -- all the dependencies showed up on the command line, including curses.h, and pdcurses.a twice. And apparently, this can mess up some old version (?) of MinGW. So, revert to spelling out "tuidemo.o tui.o". Reported by "Howard L."
Extensive documentation revision and reorganizing. More to do here. For example, I moved the build instructions from INSTALL (which never really described installation) to the platform-specific READMEs.
New indentation standard: four spaces, no tabs.
This release adds an SDL backend, refines the demos, and is faster in some cases.
SDL port. See INSTALL, doc/sdl.txt and sdl1/* for details.
Double-buffering -- minimize screen writes by checking, in doupdate() and wnoutrefresh(), whether the changes to curscr are really changes. In most cases, this makes no difference (writes were already limited to areas marked as changed), but it can greatly reduce the overhead from touchwin(). It also helps if you have small, separated updates on the same line.
The PDC_RGB colors can now be used, or not, with any platform (as long as the same options are used when compiling both the library and apps). This may help if you have apps that are hardwired to assume certain definitions.
Restored the use_default_colors() stuff from the ncurses versions of the rain and worm demos, to make them "transparent" (this is useful now, with the SDL port); added transparency to newdemo.
Added setlocale() to tuidemo, to make it easier to browse files with non-ASCII characters.
Sped up firework demo by replacing unneeded clear() and init_pair() calls.
Allow exit from ptest demo by typing 'q'.
New functions for implementors: PDC_pair_content() and PDC_init_pair() (the old pdc_atrtab stuff was arguably the last remnant of code in the pdcurses directory that was based on platform details).
Implicit wrefresh() needs to be called from wgetch() when the window's cursor position is changed, even if there are no other changes.
Set SP->audible on a per-platform basis, as was documented in IMPLEMNT, but not actually being done.
Minor tweaks for efficiency and readability, notably with wscrl().
tuidemo didn't work correctly on monochrome screens when A_COLOR was defined -- the color pair numbers appeared as the corresponding character; also, the input box was (I now realize) broken with ncurses since our 2.7, and broke more subtly with PDCurses' new implicit refresh handling; also, the path to the default file for the Browse function was a bit off.
Assume in the demos that curs_set() is always available -- there's no good test for this, and the existing tests were bogus.
Made the command-line parameter for ptest work. (If given an argument, it delays that number of milliseconds between changes, instead of waiting for a key, and automatically loops five times.)
Building the Win32 DLL with MinGW or Cygwin wouldn't work from outside the platform directory.
Building the X11 port with Cygwin required manually editing the Makefile after configuring; no longer. Reported by Warren W. Gay.
Minor tightening of configure and makefiles.
Bogus references to "ACS_BLCORNER" in the border man page. Reported by "Walrii".
slk_wlabel() was not documented.
Spelling cleanup.
Changed RCSIDs to not end with a semicolon -- avoids warnings when compiling with the -pedantic option.
Merged latin-1.txt into x11.txt.
Updated config.guess and config.sub to more recent versions.
This release mainly covers changes to the build process, along with a few structural changes.
The panel library has been folded into the main library. What this means is that you no longer need to specify "-lpanel" or equivalent when linking programs that use panel functionality with PDCurses; however, panel.lib/.a is still provided (as a copy of pdcurses.lib/.a) so that you can, optionally, build your projects with no changes. It also means that panel functionality is available with the DLL or shared library. Note that panel.h remains separate from curses.h.
Setting the PDCURSES_SRCDIR environment variable is no longer required before building, unless you want to build in a location other than the platform directory. (See INSTALL.)
MinGW and Cygwin makefiles support building DLLs, via the "DLL=Y" option. Partly due to Timofei Shatrov.
Support for the Digital Mars compiler.
Watcom makefiles now use the "loaddll" feature.
Eliminated the platform defines (DOS, WIN32, OS2, XCURSES) from curses.h, except for X11-specific SCREEN elements and functions. Dynamically-linked X11 apps built against an old version will have their red and blue swapped until rebuilt. (You can define PDC_RGB to build the library with the old color scheme, but it would also have to be defined when building any new app.) Any app that depends on PDCurses to determine the platform it's building on will have to make other arrangements.
Documentation cleanup -- added more details; removed some content that didn't apply to PDCurses; moved the doc-building tool to the doc directory; changed *.man to *.txt.
The EMX makefile now accepts "DLL=Y", builds pdcurses.dll instead of curses.dll, builds either the static library or the DLL (not both at once), and links all the demos with the DLL when building it.
In Win32, read the registry only when needed: when init_color() or color_content() is called, instead of at startup.
A few additional consts in declarations.
The Win32 compilers that build DLLs now use common .def files.
panel.h functions sorted by name, as with other .h files; curses.h is no longer included by repeated inclusions of panel.h or term.h.
Simplified Borland makefiles.
Makefile.aix.in depended on a file, xcurses.exp, that was never there. This problem was fixed as part of the change to common .def files; however, I still haven't been able to test building on AIX.
Primarily clipboard-related fixes, and special UTF-8 support.
"Force UTF-8" mode, a compile-time option to force the use of UTF-8 for multibyte strings, instead of the system locale. (Mainly for Windows, where UTF-8 doesn't work well in the console.) See INSTALL.
Multibyte string support in PDC_*clipboard() functions, and in Win32's PDC_set_title().
Added the global string "ttytype", per other curses implementations, for compatibility with old BSD curses.
Real functions for the "quasi-standard aliases" -- crmode(), nocrmode(), draino(), resetterm(), fixterm() and saveterm(). (Corresponding macros removed.)
In Win32, under NT-family OSes, the scrollback buffer would be restored by endwin(), but would not be turned off again when resuming curses after an endwin(). The result was an odd, partly-scrolled-up display. Now, the buffer is toggled by PDC_reset_prog_mode() and PDC_reset_shell_mode(), so it's properly turned off when returning from an endwin().
In 3.0, selection in X11 didn't work. (Well, the selecting worked, but the pasting elsewhere didn't.) This was due to the attempted fix "don't return selection start as a press event," so that's been reverted for now.
PDC_setclipboard() was locking up in X11. Reported by Mark Hessling.
Missing underscore in the declaration of XC_say() prevented compilation with PDCDEBUG defined. Reported by M.H.
Off-by-one error in copywin() -- the maximum coordinates for the destination window should be inclusive. Reported by Tiago Dionizio.
Start in echo mode, per X/Open. Reported by T.D.
Strip leading and trailing spaces from slk labels, per a literal reading of X/Open. Suggested by Alexey Miheev (about ncurses, but it also applies here).
The #endif for PDCURSES needs to come after the closing of the extern "C". This has been broken since June 2005. Fortunately (?), it only shows up if the file is included multiple times, and then only in C++. Reported on the DOSBox forums.
Use CF_OEMTEXT instead of CF_TEXT in the narrow versions of the clipboard functions in Win32, to match the console.
Changed the format of the string returned from longname().
In the clipboard test in the testcurs demo, use a single mvprintw() to display the return from PDC_getclipboard(), instead of a loop of addch(), which was incompatible with multibyte strings.
Moved has_key() into the keyname module, and documented it.
Moved RIPPEDOFFLINE to curspriv.h.
Typos in IMPLEMNT.
The focuses for this release are X/Open conformance, i18n, better color support, cleaner code, and more consistency across platforms.
This is only a brief summary of the changes. For more details, consult the CVS log.
An almost complete implementation of X/Open curses, including the wide-character and attr_t functions (but excluding terminfo). The wide-character functions work only in Win32 and X11, for now, and require building the library with the appropriate options (see INSTALL). Note that this is a simplistic implementation, with exactly one wchar_t per cchar_t; the only characters it handles properly are those that are one column wide.
Support for X Input Methods in the X11 port (see INSTALL). When built this way, the internal compose key support is disabled in favor of XIM's, which is a lot more complete, although you lose the box cursor.
Multibyte character support in the non-wide string handling functions, per X/Open. This only works when the library is built with wide- character support enabled.
Mouse support for DOS and OS/2. The DOS version includes untested support for scroll wheels, via the "CuteMouse" driver.
An ncurses-compatible mouse interface, which can work in parallel with the traditional PDCurses mouse interface. See the man page (or mouse.c) for details.
DOS and OS/2 can now return modifiers as keys, as in Win32 and X11.
COLORS, which had been fixed at 8, is now either 8 or 16, depending on the terminal -- usually 16. When it's 8, blinking mode is enabled (controlled as before by the A_BLINK attribute); when it's 16, bright background colors are used instead. On platforms where it can be changed, the mode is toggled by the new function PDC_set_blink(). PDCurses tries to set PDC_set_blink(FALSE) at startup. (In Win32, it's always set to FALSE; in DOS, with other than an EGA or VGA card, it can't be.) Also, COLORS is now set to 0 until start_color() is called.
Corresponding to the change in COLORS, COLOR_PAIRS is now 256.
Working init_color() and color_content(). The OS/2 version of init_color() works only in a full-screen session; the Win32 version works only in windowed mode, and only in NT-family OSes; the DOS version works only with VGA adapters (real or simulated). The Win32 version is based mostly on James Brown's setconsoleinfo.c (www.catch22.net).
use_default_colors(), assume_default_colors(), and curses_version(), after ncurses.
Added global int TABSIZE, after ncurses and Solaris curses; removed window-specific _tabsize.
Logical extension to the wide-character slk_ funcs: slk_wlabel(), for retrieving the label as a wide-character string.
A non-macro implementation of ncurses' wresize().
Working putwin(), getwin(), scr_dump() and scr_restore().
A working acs_map[]. Characters from the ACS are now stored in window structures as a regular character plus the A_ALTCHARSET attribute, and rendered to the ACS only when displayed. (This allows, for example, the correct display on one platform of windows saved from another.)
In X11, allow selection and paste of UTF8_STRING.
The testcurs demo now includes a color chart and init_color() test, a wide character input test, a display of wide ACS characters with sample Unicode text, a specific test of flash(), more info in the resize test, and attempts to change the width as well as the height.
Command-line option for MSVC to build DLLs (see INSTALL). Also, the naming distinction for DLLs ("curses" vs. "pdcurses") is abandoned, and either the static lib or DLL is built, not both at once (except for X11).
For backwards compatibility, a special module just for deprecated functions -- currently PDC_check_bios_key(), PDC_get_bios_key(), PDC_get_ctrl_break() and PDC_set_ctrl_break(). These shouldn't be used in applications, but currently are... in fact, all the "private" functions (in curspriv.h) are subject to change and should be avoided.
A new document, IMPLEMNT, describing PDCurses' internal functions for those wishing to port it to new platforms.
Mark Hessling has released the X11 port to the public domain. (However, x11/ScrollBox* retain their separate copyright and MIT-like license.)
Most of the macros have been removed (along with the NOMACROS ifdef). The only remaining ones are those which have to be macros to work, and those that are required by X/Open to be macros. There were numerous problems with the macros, and no apparent reason to keep them, except tradition -- although it was PCcurses 1.x that first omitted them.
Clean separation of platform-specific code from the rest. Outside of the platform directories, there remain only a few ifdefs in curses.h and curspriv.h.
General reorganization and simplification.
Documentation revisions.
When expanding control characters in addch() or insch(), retain the attributes from the chtype.
Preserve the A_ALTCHARSET attribute in addch() and insch().
Per X/Open, beep() should always return OK.
On platforms with a controlling terminal (i.e., not X11), curs_set(1) now sets the cursor to the shape it had at the time of initscr(), rather than always making it small. (Exception for DOS: If the video mode has been changed by PDC_resize_screen(), curs_set(1) reverts to line 6/7.) The shape is taken from SP->orig_cursor (the meaning of which is platform-specific).
Stop updating the cursor position when the cursor is invisible (this gives a huge performance boost in Win 9x); update the cursor position from curs_set() if changing from invisible to visible.
Some tweaking of the behavior of def_prog_mode(), def_shell_mode(), savetty(), reset_prog_mode(), reset_shell_mode() and resetty()... still not quite right.
flash() was not implemented for Win32 or X. A portable implementation is now used for all platforms. Note that it's much slower than the old (DOS and OS/2) version, but this is only apparent on an extremely slow machine, such as an XT.
In getstr(), backspacing on high-bit characters caused a double backspace.
hline() and vline() used an incorrect (off by one) interpretation of _maxx and _maxy. If values of n greater than the max were specified, these functions could access unallocated memory.
innstr() is supposed to return the number of characters read, not just OK or ERR. Reported by Mike Aubury.
A proper implementation of insch() -- the PDC_chadd()-based version wasn't handling the control characters correctly.
Return ASCII and control key names from keyname() (problem revealed by ncurses' movewindow test); also, per X/Open, return "UNKNOWN KEY" when appropriate, rather than "NO KEY NAME".
Turn off the cursor from leaveok(TRUE), even in X11; leaveok(FALSE) now calls curs_set(1), regardless of the previous state of the cursor.
In the slk area, BUTTON_CLICKED events now translate to function keys, along with the previously recognized BUTTON_PRESSED events. Of course, it should really be checking the events specified by map_button(), which still doesn't work.
napms(0) now returns immediately.
A unified napms() implementation for DOS -- no longer throttles the CPU when built with any compiler.
Allow backspace editing of the nocbreak() buffer.
pair_content(0, ...) is valid.
There was no check to ensure that the pnoutrefresh() window fit within the screen. It now returns an ERR if it doesn't.
In X11, resize_term() must be called with parameters (0, 0), and only when SP->resized is set, else it returns ERR.
Copy _bkgd in resize_window(). Patch found on Frederic L. W. Meunier's web site.
slk_clear() now removes the buttons completely, as in ncurses.
Use the current foreground color for the line attributes (underline, left, right), unless PDC_set_line_color() is explicitly called. After setting the line color, you can reset it to this mode via "PDC_set_line_color(-1)".
Removed non-macro implementations of COLOR_PAIR() and PAIR_NUMBER().
Dispensed with PDC_chadd() and PDC_chins() -- waddch() and winsch() are now (again) the core functions.
Dropped or made static many obsolete, unused, and/or broken functions, including PDC_chg_attrs(), PDC_cursor_on() and _off(), PDC_fix_cursor(), PDC_get_attribute(), PDC_get_cur_col() and _row(), PDC_set_80x25(), PDC_set_cursor_mode(), PDC_set_rows(), PDC_wunderline(), PDC_wleftline(), PDC_wrightline(), XCursesModifierPress() and XCurses_refresh_scrollbar().
Obsolete/unused defines: _BCHAR, _GOCHAR, _STOPCHAR, _PRINTCHAR _ENDLINE, _FULLWIN and _SCROLLWIN.
Obsolete/unused elements of the WINDOW struct: _pmax*, _lastp*, _lasts*.
Obsolete/unused elements of the SCREEN struct: orgcbr, visible_cursor, sizeable, shell, blank, cursor, orig_emulation, font, orig_font, tahead, adapter, scrnmode, kbdinfo, direct_video, video_page, video_seg, video_ofs, bogus_adapter. (Some of these persist outside the SCREEN struct, in the platform directories.) Added mouse_wait and key_code.
Removed all the EMALLOC stuff. Straight malloc calls were used elsewhere; it was undocumented outside of comments in curspriv.h; and there are better ways to use a substitute malloc().
Single mouse clicks are now reportable on all platforms (not just double-clicks). And in general, mouse event reporting is more consistent across platforms.
The mouse cursor no longer appears in full-screen mode in Win32 unless a nonzero mouse event mask is used.
ALT-keypad input now works in Win32.
In Win32, SetConsoleMode(ENABLE_WINDOW_INPUT) is not useful, and appears to be the source of a four-year-old bug report (hanging in THE) by Phil Smith.
Removed the PDC_THREAD_BUILD stuff, which has never worked. For the record: PDCurses is not thread-safe. Neither is ncurses; and the X/Open curses spec explicitly makes it a non-requirement.
With the internal compose key system in the X11 port, modifier keys were breaking out of the compose state, making it impossible to type accented capitals, etc. Also, Multi_key is now the default compose key, instead of leaving it undefined by default; and a few more combos are supported.
In X11, the first reported mouse event after startup always read as a double-click at position 0, 0. (This bug was introduced in 2.8.)
In X11, don't return selection start as a press event. (Shift-click on button 1 is still returned.)
In X11, properly handle pasting of high-bit chars. (It was doing an unwanted sign extension.)
In X11, BUTTON_MOVED was never returned, although PDC_MOUSE_MOVED was set.
The fix in 2.8 for the scroll wheel in X11 wasn't very good -- it did report the events as scroll wheel events, but it doubled them. Here's a proper fix.
Changed mouse handling in X11: Simpler translation table, with XCursesPasteSelection() called from XCursesButton() instead of the translation table; require shift with button 1 or 2 for select or paste when mouse events are being reported (as with ncurses), allowing passthrough of simple button 2 events. This fixes the previously unreliable button 2 behavior.
Modifier keys are now returned on key up in X11, as in Win32. And in general, modifier key reporting is more consistent across platforms.
Modifiers are not returned as keys when a mouse click has occurred since the key press.
In BIOS mode (in DOS), count successive identical output bytes, and make only one BIOS call for all of them. This dramatically improves performance.
The cursor position was not always updated correctly in BIOS mode.
In testcurs, the way the ACS test was written, it would really only work with a) PDCurses (with any compiler), or b) gcc (with any curses). Here's a more portable implementation.
Better reporting of mouse events in testcurs.
Blank out buffer and num before the scanw() test in testcurs, in case the user just hits enter or etc.; clear the screen after resizing.
Allow tuidemo to use the last line.
Separate left/right modifier keys are now reported properly in Win32. (Everything was being reported as _R.)
Attempts to redirect input in Win32 now cause program exit and an error message, instead of hanging.
Dropped support for the Microway NDP compiler.
Some modules renamed, rearranged.
Fixes for errors and warnings when building with Visual C++ 2005.
In MSVC, the panel library didn't work with the DLL.
Complete export lists for DLLs.
Simplified makefiles; moved common elements to .mif files; better optimization; strip demos when possible.
Changed makefile targets of "pdcurses.a/lib" and "panel.a/lib" to $(LIBCURSES) and $(LIBPANEL). Suggestion of Doug Kaufman.
Changed "install" target in the makefile to a double-colon rule, to get around a conflict with INSTALL on non-case-sensitive filesystems, such as Mac OS X's HFS+. Reported by Douglas Godfrey et al.
Make PDCurses.man dependent on manext. Suggestion of Tiziano Mueller.
Set up configure.ac so autoheader works; removed some obsolescent macros. Partly the suggestion of T.M.
The X11 port now builds in the x11 directory (including the demos), as with other ports.
The X11 port should now build on more 64-bit systems. Partly due to M.H.
The default window title and icons for the X11 port are now "PDCurses" instead of "XCurses".
Internal functions and variables made static where possible.
Adopted a somewhat more consistent naming style: Internal functions with external linkage, and only those, have the prefix "PDC_"; external variables that aren't part of the API use "pdc_"; static functions use ""; and "XC" and "xc_" prefixes are used for functions and variables, respectively, that are shared between both processes in the X11 port. Also eliminated camel casing, where possible.
Changed the encoding for non-ASCII characters in comments and documentation from Latin-1 to UTF-8.
As with the previous version, you should assume that apps linked against older dynamic versions of the library won't work with this one until recompiled.
New features:
Simpler, faster.
Declarations for all supported, standard functions, per the X/Open Curses 4.2 spec, with the notable exception of getch() and ungetch(). You can disable the use of the macro versions by defining NOMACROS before including curses.h (see xmas.c for an example). NOMACROS yields smaller but theoretically slower executables.
New functions: vwprintw(), vwscanw(), vw_printw() and vw_scanw(). This completes the list of X/Open 4.2 functions, except for those concerned with attr_t and wide characters. Some (especially the terminfo/termcap functions) aren't yet fully fleshed out, though.
Non-macro implementations for COLOR_PAIR(), PAIR_NUMBER(), getbkgd(), mvgetnstr(), mvwgetnstr(), mvhline(), mvvline(), mvwhline(), and mvwvline(). (The macros are still available, too.)
newterm() works now, in a limited way -- the parameters are ignored, and only the first invocation will work (i.e., only one SCREEN can be used).
start_color() works now -- which is to say, if you don't call it, you'll only get monochrome output. Also, without calling it, the terminal's default colors will be used, where supported (currently only in Win32). This is equivalent to the PDC_ORIGINAL_COLORS behavior introduced in 2.7, except that only the default colors will be used. (PDC_ORIGINAL_COLORS is still available, if you want to combine the use of specific colors and the default colors.)
New logic for termname() and longname(): termname() always returns "pdcurses"; longname() returns "PDCurses for [platform] [adapter] [COLOR/MONO]-YxX" (adapter is only defined for DOS and OS/2). This is the first time these functions return anything in Win32.
New installation method for XCurses: the header files are placed in a subdirectory "xcurses" within the include directory, rather than being renamed. (But the renamed xcurses.h and xpanel.h are also installed, for backwards compatibility.) curspriv.h and term.h are now available, and existing curses-based code need no longer be edited to use XCurses' curses.h. And with no more need for explicit XCursesExit() calls (see below), your code need not be changed at all to move from another curses implementation to XCurses. It can be as simple as "gcc -I/usr/local/include/xcurses -lXCurses -oprogname progname.c".
Combined readme.* into this HISTORY file, and incorporated the old 1.x (PCcurses) history.
New functionality for the testcurs demo: ACS character display; menu support for PgUp, PgDn, Home and End; centered menu; and it can now be resized in X.
Added modified versions of the rain and worm demos from ncurses.
Bug fixes and such:
Big cleanup of dead and redundant code, including unneeded defines, ifdefs, and structure elements.
flushinp() was not implemented for Win32.
resetty() was not restoring LINES and COLS.
nonl() made '\n' print a line feed without carriage return. This was incorrect.
Removed bogus implementation of intrflush().
The line-breakout optimization system, disabled by default in 2.7, is removed in 2.8. It simply didn't work, and never has. (The typeahead() function remains, for compatibility, but does nothing.)
The declarations for the printw() and scanw() function families were erroneously ifdef'd.
Safer printw() calls on platforms that support vsnprintf().
Use the native vsscanf() in DJGPP, MinGW and Cygwin.
ACS_BLOCK now works in X.
Explicit calls to XCursesExit() are no longer needed.
XCURSES is now defined automatically if not DOS, OS2 or WIN32.
The default icon for XCurses wasn't working (had to remove the focus hint code to fix this). Also, the default title is now "XCurses" instead of "main".
Incorrect dimensions (undercounting by two in each direction) were shown while resizing in X.
Scroll wheel events were not always correctly reported in X.
32 bits are enough for the "long" chtype, but 64 bits were used on a 64-bit system, wasting memory. Now conditioned on _LP64. This could be faster, too.
The short, 16-bit chtype now works with XCurses.
Corrected return value for is_linetouched(), is_wintouched(), can_change_color() and isendwin() (bool instead of int).
timeout(), wtimeout(), idcok() and immedok() return void.
pair_content() takes a short.
Replaced incorrect usages of attr_t with chtype. attr_t is still typedef'd, for backwards compatibility. (It's supposed to be used for the WA_*-style functions, which PDCurses doesn't yet support.)
Added const where required by the spec, and in other appropriate places.
Removed PDC_usleep(). napms() is now the core delay routine.
Fixed poll() support in napms().
Various changes to the internal PDC_* functions -- don't depend on these, and don't use them unless you absolutely have to.
Some routines accessed window structures in their variable declarations, before checking for a NULL window pointer.
Dropped support for the undocumented PDC_FULL_DISPLAY, wtitle(), and PDC_print().
Cleaned up remaining warnings.
Reduced unnecessary #include directives -- speeds up compilation.
Fix for demos build in Borland/DOS -- the makefile in 2.7 didn't specify the memory model. Reported by Erwin Waterlander.
Simplified the makefiles; e.g., some now build each demo in a single step, and Watcom no longer uses demos.lnk. Also, the demo exes are now stripped when possible; maximum compression used for archives built by the makefiles; xcurses-config removed as part of "make distclean"; and I tweaked optimization for some platforms.
Reverted to /usr/local/ as default installation directory for XCurses.
Upgraded to autoconf 2.59... instantly doubling the size of the configure script. Ah well. Otherwise, simplified the build system.
Dropped support for pre-ANSI compilers. (It hasn't worked since at least version 2.4, anyway.)
Revised and, I hope, clarified the boilerplate and other comments.
Simplified logging and RCS ids; added RCS ids where missing.
Consistent formatting for all code, approximately equivalent to "indent -kr -i8 -bl -bli0", with adjustments for 80 columns.
INTRODUCTION:
Hello all. As of a few weeks ago, I'm the new maintainer for PDCurses. Here's a brief summary of changes in this release. (More details are available in the CVS log and trackers on SourceForge.)
NEW FEATURES:
Functions: delscreen(), getattrs(), has_key(), slk_color(), wcolor_set(), wtimeout().
Macros: color_set(), mvhline(), mvvline(), mvwgetnstr(), mvwhline(), mvwvline(), timeout(), wresize().
Stub implementations of terminfo functions (including a term.h).
More stubs for compatibility: filter(), getwin(), putwin(), noqiflush(), qiflush(), scr_dump(), scr_init(), scr_restore(), scr_set(), use_env(), vidattr(), vidputs().
The terminal's default colors are used as curses' default colors when the environment variable "PDC_ORIGINAL_COLORS" is set to any value (Win32 only at the moment).
Simplified build system.
Replaced PDC_STATIC_BUILD with its opposite, PDC_DLL_BUILD (see .mak files for more info).
Minimal implementation of color_content() -- no longer a stub.
Added the remaining ACS defines (ACS_S3, ACS_BBSS, etc.) for DOS/OS2/Win; "enhanced" versions of existing ACS characters used.
Support for scroll wheels.
Support for Pacific C.
BUGS FIXED:
Builds correctly (including demos) on all tested platforms (see below); nearly all compiler warnings have been cleaned up; the ptest demo is built on all platforms; "clean" targets are improved.
The ability to build ncurses_tests has been restored (see demos dir).
Line-breakout optimization now defaults to off (equivalent to "typeahead(-1)"), so output is not interrupted by keystrokes (it's supposed to resume on the next refresh(), which wasn't working).
Implicit wrefresh() in wgetch() was not being invoked in nodelay mode.
subpad() was erroneously offsetting from the origin coordinates of the parent pad (which are always -1,-1).
In wborder(), whline(), and wvline(), the current (wattrset) attribute was being used, but not the current background (wbkgd).
Allow Russian 'r' character ASCII 0xe0 to be returned.
termattrs() now also returns A_UNDERLINE, A_REVERSE.
In Win32, with large scrollback buffers set, there was an unwanted "scrollup" effect on startup.
Revamped keyboard handling for Win32.
New screen resize method for Win32.
napms(), delay_output(), etc. now work with Cygwin.
curs_set(0) wasn't working in Win32 in full-screen (ALT-ENTER) mode -- the cursor stayed on.
The A_REVERSE attribute was broken in XCurses.
On 64-bit systems, XCurses was ignoring every other keystroke.
Added focus hints for XCurses.
Demos (except for tuidemo) once again have their proper titles in XCurses (using Xinitscr() instead of the obsolete XCursesProgramName).
The 16-bit chtype is a working option again (by removing #define CHTYPE_LONG from curses.h), except in XCurses. It's not recommended; but if your needs are limited, it still works.
Reset screen size in resetty() under DOS, as in Win32 and OS/2.
Changes for cursor size under DOS.
Automatic setting of BIOS mode for CGA under DOS now works.
The cursor is now always updated in PDC_gotoxy(); this fixes the problem of missing characters in BIOS mode.
Macros nocbreak(), cbreak(), nocrmode(), crmode(), nodelay(), nl() and nonl() now return OK.
ERR and OK are now defined as -1 and 0, respectively, for compatibility with other curses implementations -- note that this change is not binary compatible; you'll have to rebuild programs that use shared/dynamic libraries.
Added "const" to prototypes where appropriate.
Miscellaneous code cleanup.
ACKNOWLEDGEMENTS:
and of course, MARK HESSLING, for his over 13 years of service as the maintainer of PDCurses. Plus, thanks to all who've reported bugs or requested features. Apologies to anyone I've forgotten.
I've tested this version on Turbo C++ 3.0 and Borland C++ 3.1 for DOS; DJGPP 2.X; Open Watcom 1.3 for DOS (16 and 32-bit), Windows and OS/2; EMX 0.9d and the "newgcc" version of EMX; Borland C++ 5.5 for Windows; recent versions of MinGW, Cygwin, LCC-Win32 and Microsoft Visual C++; and gcc under several flavors of Linux, Mac OS X, *BSD and Solaris.
-- William McBrine
INTRODUCTION:
This release of PDCurses includes the following changes:
BUGS FIXED:
Allow accented characters on Win32 platform when run on non-English keyboards.
Allow "special" characters like Ctrl-S, Ctrl-Q under OS/2 to be returned.
Some bugs with halfdelay() fixed by William McBrine.
pechochar() should now work correctly.
redrawwin() macro in curses.h was incorrect - fixed by Alberto Ornaghi
Don't include "special" characters like KEY_SHIFT_L to be returned in getnstr() family. Bug 542913
Entering TAB in wgetnstr() no longer exceeds requested buffer size. Bug 489233
Fixed bug 550066, scrollok() and pads. Also beep() called when buffer exceeded. Bug 562041.
Reverse video of X11 selection reinstated. Pablo Garcia Abio??
Right Alt modifier now works like left Alt modifier under Win32
Add support for all libXaw replacement libraries with Scrollbar bug. Note that for this to work, you still have to change the libXaw replacement libraries to fix the bug :-(
Don't trap signals in XCurses if calling application has ignored them. Change by Frank Heckenbach.
Bug reports from Warren W. Gay:
Fixed bug with calling def_prog_mode(), resize_term(), reset_prog_mode(); the resize details were being lost.
NEW FEATURES:
Clipboard support now available on DOS platform, but handled internally to the currently running process.
New X11 resource: textCursor, allows the text cursor to be specified as a vertical bar, or the standard horizontal bar. Thanks to Frank Heckenbach for the suggestion.
NEW COMPILER SUPPORT:
INTRODUCTION:
This release of PDCurses includes the following changes:
Set BASE address for Win32 DLL
Add KEY_SUP and KEY_SDOWN.
Add PDC_set_line_color()
Add blink support as bold background
Add bold colors
Add getbkgd() macro
Add new PDC functions for adding underline, overline, leftline and rightline
Add support for shifted keypad keys.
Allow more keypad keys to work under Win32
Change Win32 and OS/2 DLL name to curses.dll
Change example resources to allow overriding from the command line
Changes for building cleanly on OS/2
Changes to handle building XCurses under AIX
Check if prefresh() and pnoutrefresh() parameters are valid.
Ensure build/install works from any directory
Handle platforms where X11 headers do not typedef XPointer.
Mention that Flexos is likely out-of-date.
Pass delaytenths to XCurses_rawgetch()
Remove boldFont
Updates for cursor blinking and italic.
BUGS FIXED:
Fix bug with getting Win32 clipboard contents. Added new PDC_freeclipboard() function.
Fix bug with halfdelay()
Fix bug with mouse interrupting programs that are not trapping mouse events under Win32.
Fix return value from curs_set()
Reverse the left and right pointing bars in ALT_CHARSET
NEW COMPILER SUPPORT:
INTRODUCTION:
This release of PDCurses includes the following changes:
full support of X11 selection handling
removed the need for the cursos2.h file
enabled the "shifted" key on the numeric keypad
added native clipboard support for X11, Win32 and OS/2
added extra functions for obtaining internal PDCurses status
added clipboard and key modifier tests in testcurs.c
fixes for panel library
key modifiers pressed by themselves are now returned as keys: KEY_SHIFT_L KEY_SHIFT_R KEY_CONTROL_L KEY_CONTROL_R KEY_ALT_L KEY_ALT_R This works on Win32 and X11 ports only
Added X11 shared library support
Added extra slk formats supported by ncurses
Fixed bug with resizing the terminal when slk were on.
Changed behavior of slk_attrset(), slk_attron() slk_attroff() functions to work more like ncurses.
BUGS FIXED:
NEW FUNCTIONS:
PDC_getclipboard() and PDC_setclipboard() for accessing the native clipboard (X11, Win32 and OS/2)
PDC_set_title() for setting the title of the window (X11 and Win32 only)
PDC_get_input_fd() for getting the file handle of the PDCurses input
PDC_get_key_modifiers() for getting the keyboard modifier settings at the time of the last (w)getch()
Xinitscr() (only for X11 port) which allows standard X11 switches to be passed to the application
NEW COMPILER SUPPORT:
MingW32 GNU compiler under Win95/NT
Cygnus Win32 GNU compiler under Win95/NT
Borland C++ for OS/2 1.0+
lcc-win32 compiler under Win95/NT
ACKNOWLEDGEMENTS: (for this release)
INTRODUCTION:
This release of PDCurses includes the following changes:
added more System V R4 functions
added Win32 port
the X11 port is now fully functional
the MS Visual C++ Win32 port now includes a DLL
both the X11 and Win32 ports support the mouse
the slk..() functions are now functional
support for scrollbars under X11 are experimental at this stage
long chtype extended to non-Unix ports
The name of the statically built library is pdcurses.lib (or pdcurses.a). The name of the DLL import library (where applicable) is curses.lib.
BUGS FIXED:
NEW FUNCTIONS:
NEW COMPILER SUPPORT:
MS Visual C++ under Win95/NT
Watcom C++ under OS/2, Win32 and DOS
two EMX ports have been provided:
EXTRA OPTIONS:
PDCurses recognizes two environment variables which determines the initialization and finalization behavior. These environment variables do not apply to the X11 port.
PDC_PRESERVE_SCREEN - If this environment variable is set, PDCurses will not clear the screen to the default white on black on startup. This allows you to overlay a window over the top of the existing screen background.
PDC_RESTORE_SCREEN - If this environment variable is set, PDCurses will take a copy of the contents of the screen at the time that PDCurses is started; initscr(), and when endwin() is called, the screen will be restored.
ACKNOWLEDGEMENTS: (for this release)
INTRODUCTION:
This release of PDCurses has includes a number of major changes:
The portable library functions are now grouped together into single files with the same arrangement as System V R4 curses.
A panels library has been included. This panels library was written by Warren Tucker.
Quite a few more functions have been supplied by Wade Schauer and incorporated into release 2.2. Wade also supplied the support for the Microway NDP C/C++ 32 bit DOS compiler.
The curses datatype has been changed from an unsigned int to a long. This allows more attributes to be stored as well as increasing the number of color-pairs from 32 to 64.
Xwindows port (experimental at the moment).
BUGS FIXED:
mvwin() checked the wrong coordinates
removed DESQview shadow memory buffer checking bug in curses.h in #define for wstandout()
lots of others I can't remember
NEW FUNCTIONS:
COMPILER SUPPORT:
DJGPP 1.12 is now supported. The run-time error that caused programs to crash has been removed.
emx 0.9a is supported. A program compiled for OS/2 should also work under DOS if you use the VID=EMX switch when compiling. See the makefile for details.
The Microway NDP C/C++ DOS compiler is now supported. Thanks to Wade Schauer for this port.
The Watcom C++ 10.0 DOS compiler is now supported. Thanks to Pieter Kunst for this port.
The library now has many functions grouped together to reduce the size of the library and to improve the speed of compilation.
The "names" of a couple of the compilers in the makefile has changed; CSET2 is now ICC and GO32 is now GCC.
EXTRA OPTIONS:
One difference between the behavior of PDCurses and Unix curses is the attributes that are displayed when a character is cleared. Under Unix curses, no attributes are displayed, so the result is always black. Under PDCurses, these functions clear with the current attributes in effect at the time. With the introduction of the bkgd functions, by default, PDCurses clears using the value set by (w)bkgd(). To have PDCurses behave the same way as it did before release 2.2, compile with -DPDCURSES_WCLR
ACKNOWLEDGEMENTS: (for this release)
Pieter Kunst, David Nugent, Warren Tucker, Darin Haugen, Stefan Strack, Wade Schauer and others who either alerted me to bugs or supplied fixes.
INTRODUCTION:
The current code contains bug fixes for the DOS and OS/2 releases and also includes an alpha release for Unix. The Unix release uses another public domain package (mytinfo) to handle the low-level screen writes. mytinfo was posted to comp.sources.unix (or misc) in December 1992 or January 1993. Unless you are a glutton for punishment I would recommend you avoid the Unix port at this stage.
The other major addition to PDCurses is the support for DJGPP (the DOS port of GNU C++). Thanks to David Nugent davidn@csource.oz.au.
Other additions are copywin() function, function debugging support and getting the small and medium memory models to work. The testcurs.c demo program has also been changed significantly and a new demo program, tuidemo, has been added.
Some people have suggested including information on where to get dmake from. oak.oakland.edu in /pub/msdos/c
OTHER NOTES:
Under DOS, by default, screen writes to a CGA monitor are done via the video BIOS rather than by direct video memory writes. This is due to the CGA "snow" problem. If you have a CGA monitor and do not suffer from snow, you can compile private_queryad.c with CGA_DIRECT defined. This will then use cause PDCurses to write directly to the CGA video memory.
Function debugging: Firstly to get function debugging, you have to compile the library with OPT=N in the makefile. This also turns on compiler debugging. You can control when you want PDCurses to write to the debug file (called trace in the current directory) by using the functions traceon() and traceoff() in your program.
With the addition of several new functions, using dmake to compile PDCurses now causes the compiler to run "out of heap space in pass 2". Using the 6.00AX version (DOS-Extended) to compile PDCurses fixes this problem; hence the -EM switch.
Added OS/2 DLL support.
A few curses functions have been fixed to exhibit their correct behavior and make them more functionally portable with System V curses. The functions that have changed are overlay(), overwrite() and typeahead.
overlay() and overwrite()
Both of theses functions in PDCurses 2.0 allowed for one window to be effectively placed on top of another, and the characters in the first window were overlaid or overwritten starting at 0,0 in both windows. This behavior of these functions was not correct. These functions only operate on windows that physically overlap with respect to the displayed screen. To achieve the same functionality as before, use the new function copywin(). See the manual page for further details.
typeahead()
This function in PDCurses 2.0 effectively checked to see if there were any characters remaining in the keyboard buffer. This is not the behavior exhibited by System V curses. This function is intended purely to set a flag so that curses can check while updating the physical screen if any keyboard input is pending. To achieve the same effect with typeahead() under PDCurses 2.1 the following code should be used.
In place of...
while(!typeahead(stdin)) { /* do something until any key is pressed... */ }
use...
/* getch() to return ERR if no key pending / nodelay(stdscr,TRUE); while(getch() == (ERR)) { / do something until any key is pressed... */ }
ACKNOWLEDGEMENTS: (in no particular order)
Jason Shumate, Pieter Kunst, David Nugent, Andreas Otte, Pasi Hamalainen, James McLennan, Duane Paulson, Ib Hojme
Apologies to anyone I may have left out.
INTRODUCTION:
Well, here it finally is; PDCurses v2.0.
PDCurses v2.0 is an almost total rewrite of PCcurses 1.4 done by John 'Frotz' Fa'atuai, the previous maintainer. It adds support for OS/2 as well as DOS.
This version has been tested with Microsoft C v6.0, QuickC v2.0 and Borland C++ 2.0 under DOS and Microsoft C v6.0 and TopSpeed c v3.02 under OS/2 2.0. Also the library has been compiled successfully with emx 0.8e, C Set/2 and Watcom 9. Most testing was done with the large memory model, where applicable. The large memory model is probably the best model to use.
The amount of testing has not been as extensive as I would have liked, but demands on releasing a product have outweighed the product's quality. Nothing new with that !! Hopefully with wider circulation, more bugs will be fixed more quickly.
I have included just 1 makefile which is suitable for dmake 3.8 for both DOS and OS/2. The makefile does not rely on customization of the dmake.ini file.
If you discover bugs, and especially if you have fixes, please let me know ASAP.
The source to the library is distributed as a zip file made with zip 1.9. You will need Info-ZIP unzip 5.0 to unzip. Follow the directions below to compile the library.
DIRECTIONS:
Create a new directory in which to unzip pdcurs20.zip. This will create a curses directory and a number of subdirectories containing source code for the library and utilities and the documentation.
Make changes to the makefile where necessary: Change the MODEL or model macro to the appropriate value (if it applies to your compiler). Use model for Borland compilers.
Change any paths in the defined macros to be suitable for your compiler.
Invoke DMAKE [-e environment_options] [target]
where environment_options are:
OS (host operating system)
COMP (compiler)
OPT (optimized version or debug version) - optional. default Y
TOS (target operating system) - optional. default OS
see the makefile for valid combinations
targets: all, demos, lcursesd.lib, manual...
NB. dmake is case sensitive with targets, so those environments that use an upper case model value (eg MSC) MUST specify the library target as for eg. Lcursesd.lib
The makefile is by default set up for Borland C++. The use of -e environment_options override these defaults. If you prefer, you can just change the defaults in the makefile and invoke it without the -e switch.
OTHER NOTES:
The documentation for the library is built into each source file, a couple of specific doc files and the header files. A program is supplied (manext) to build the manual. This program gets compiled when you build the documentation.
To generate the library response file correctly, I had to write a quick and dirty program (buildlrf) to achieve this. Originally the makefiles just had statements like: "echo -+$(OBJ)$* & >> $(LRF)" which appended a suitable line to the response file. Unfortunately under some combinations of makefiles and command processors (eg. nmake and 4DOS) the & would get treated as stderr and the echo command would fail.
The original source for PDCurses that I received from the previous maintainer contained support for the FLEXOS operating system. Not having access to it, I could not test the changes I made so its support has fallen by the wayside. If you really need to have PDCurses running under FLEXOS, contact me and I will see what can be arranged.
Under DOS, by default, screen writes to a CGA monitor are done via the video BIOS rather than by direct video memory writes. This is due to the CGA "snow" problem. If you have a CGA monitor and do not suffer from snow, you can compile private_queryad.c with CGA_DIRECT defined. This will then use cause PDCurses to write directly to the CGA video memory.
Added System V color support.
COMPILER-SPECIFIC NOTES:
It is possible with MSC 6.0 to build the OS/2 libraries and demo programs from within DOS. This is the only case where it is possible to specify the value of TOS on the command line to be OS2 and the value of OS be DOS.
I have only tested the library using the migration libraries. I doubt that the demo programs will work without them.
Testing has been done with 0.8e of emx together with the 16_to_32 libraries. The emx\lib directory should include the vio32.lib and kbd32.lib libraries from the 16_to_32 package.
BUGS and UNFINISHED BUSINESS:
PDC_set_ctrl_break() function does not work under OS/2.
win_print() and PDC_print() do not work under OS/2.
The file todo.man in the doc directory also lists those functions of System V 3.2 curses not yet implemented. Any volunteers?
ACKNOWLEDGEMENTS:
-- Mark Hessling
Changed back from short to int. (int is the correct size for the default platform. Short might be too short on some platforms. This is more portable. I, also, made this mistake.)
Many functions are now macros. If you want the real thing, #undef the macro. (X/Open requirement.)
Merged many sources into current release.
Added many X/Open routines (not quite all yet).
Added internal documentation to all routines.
Added a HISTORY file to the environment.
Added a CONTRIB file to the environment.
Added many levels of compiler support. Added mixed prototypes for all "internal" routines. Removed all assembly language. Added EGA/VGA support. Converted all #ifdef to #if in all modules except CURSES.H and CURSPRIV.H. Always include ASSERT.H. Added support for an external malloc(), calloc() and free(). Added support for FAST_VIDEO (direct-memory writes). Added various memory model support (for FAST_VIDEO). Added much of the December 1988 X/Open Curses specification.
-- John 'Frotz' Fa'atuai
In PCcurses v.1.4, both portability improvements and bugfixes have been made. The files have been changed to allow lint-free compilation with Microsoft C v.5.1, and with Turbo C v.2.0. The source should still compile without problems on older compilers, although this has not been verified.
The makefiles have been changed to suit both the public release and the author, who maintains a special kind of libraries for himself. In the case of Microsoft C, changes were done in the makefile to lower the warning level to 2 (was 3). This was to avoid ANSI warnings which are abundant because PCcurses does not attempt to follow strict ANSI C standard.
BUG FIXES FROM V.1.3 TO V.1.4:
!!!IMPORTANT CHANGE!!!
The definitions for OK and ERR in curses.h were exchanged. This was done to be more consistent with UNIX versions. Also, it permits functions like newwin() and subwin() to return 0 (=NULL) when they fail due to memory shortage. This incompatibility with UNIX curses was pointed out by Fred C. Smith. If you have tested success/failure by comparisons to anything other than ERR and OK, your applications will need to be be changed on that point. Sorry... but presumably most of you used the symbolic constants?
(END OF IMPORTANT CHANGE)
Fred also pointed out a bug in the file update.c. The bug caused the first character printed after 'unauthorized' screen changes (like during a shell escape, for example) to be placed at the wrong screen position. This happened even if the normal precautions (clear / touch / refresh) were taken. The problem has now been fixed.
PCcurses is currently also being used on a 68000 system with hard-coded ESCape sequences for ANSI terminals. However, ints used by the 68000 C compiler are 32 bits. Therefore ints have been turned into shorts wherever possible in the code (otherwise all window structures occupy twice as much space as required on the 68000). This does not affect PC versions since normally both ints and shorts are 16 bits for PC C compilers.
At some places in the source code there are references made to the 68000 version. There are also a makefile, a curses68.c file, and a curses68.cmd file. These are for making, low-level I/O, and linking commands when building the 68000 version. These files are probably useful to no-one but the author, since it is very specific for its special hardware environment. Still in an effort to keep all curses-related sources in one place they are included. Note however that PCcurses will not officially support a non-PC environment.
The file cursesio.c, which was included in the package at revision level 1.2, and which was to be an alternative to the cursesio.asm file, has been verified to behave incorrectly in the function _curseskeytst(). The problem was that the value of 'cflag' does not contain the proper data for the test that is attempted. Furthermore, neither Turbo C or Microsoft C allows any way to return the data that is needed, and consequently you should not use cursesio.c. The best solution is to simply use the ASM version. In v.1.2 and v.1.3, the user could edit the makefile to select which version he wanted to use. The makefiles in v.1.4 have removed this possibility forcing the use of the ASM file, and cursesio.c has been dropped from the distribution.
A bug in the wgetstr() function caused PCcurses to echo characters when reading a keyboard string, even if the echo had been turned off. Thanks to Per Foreby at Lund University, Sweden, for this. Per also reported bugs concerning the handling of characters with bit 8 set. Their ASCII code were considered as lower than 32, so they were erased etc. like control characters, i.e. erasing two character positions. The control character test was changed to cope with this.
The overlay() and overwrite() functions were changed so that the overlaying window is positioned at its 'own' coordinates inside the underlying window (it used to be at the underlying window's [0,0] position). There is some controversy about this - the documentation for different curses versions say different things. I think the choice made is the most reasonable.
The border() and wborder() functions were changed to actually draw a border, since this seems to be the correct behavior of these functions. They used to just set the border characters to be used by box(). These functions are not present in standard BSD UNIX curses.
The subwin() function previously did not allow the subwindow to be as big as the original window in which it was created. This has now been fixed. There was also the problem that the default size (set by specifying numlines or numcols (or both) as 0 made the resulting actual size 1 line/column too small.
There were a few spelling errors in function names, both in the function declarations and in curses.h. This was reported by Carlos Amaral at INESC in Portugal. Thanks! There was also an unnecessary (but harmless) parameter in a function call at one place.
The file 'border.c' is now included. It allows you to explicitly specify what characters should be used as box borders when the box() functions are called. If the new border characters are non-0, they override the border characters specified in the box() call. In my understanding, this functionality is required for AT&T UNIX sV.3 compatibility. Thanks for this goes to Tony L. Hansen (hansen@pegasus.UUCP) for posting an article about it on Usenet (newsgroup comp.unix.questions; his posting was not related at all to PCcurses).
The only other difference between v.1.2 and v.1.3 is that the latter has been changed to avoid warning diagnostics if the source files are compiled with warning switches on (for Microsoft this means '-W3', for Turbo C it means '-w -w-pro'). Of these, the Turbo C warning check is clearly to be used rather than Microsoft, even if neither of them comes even close to a real UNIX 'lint'. Some of the warnings in fact indicated real bugs, mostly functions that did not return correct return values or types.
The makefiles for both MSC and TRC have been modified to produce warning messages as part of normal compilation.
The changes from v.1.1 to v.1.2 are minor. The biggest change is that there was a bug related to limiting the cursor movement if the application tried to move it outside the screen (something that should not be done anyway). Such erroneous application behavior is now handled appropriately.
All modules have been changed to have a revision string in them, which makes it easier to determine what version is linked into a program (or what library version you have).
There is now a 'cursesio.c' file. That file does the same as 'cursesio.asm' (i.e. it provides the interface to the lower-level system I/O routines). It is written in C and thus it is (possibly) more portable than the assembler version (but still not so portable since it uses 8086 INT XX calls directly). When one creates new curses libraries, one chooses whether to use the assembler or the C version of cursesio. The choice is made by commenting out the appropriate dependencies for cursesio.obj, near the end of the makefiles.
There is now a 'setmode.c' file. That file contains functions that save and restore terminal modes. They do it into other variables than do savetty() and resetty(), so one should probably use either savetty()/resetty() or the new functions only - and not mix the both ways unless one really knows what one does.
Diff lists vs v.1.0 are no longer included in the distribution. The make utility still is. PCcurses v.1.2 still compiles with Microsoft C v.4.0, and with Borland Turbo C v.1.0. There is as far as I know no reason to believe that it does not compile under Microsoft C v.3.0 and 5.x, or Turbo C v.1.5, but this has not been tested.
There are two makefiles included, one for Microsoft C, one for Turbo C. They are both copies of my personal makefiles, and as such they reflect the directory structure on my own computer. This will have to be changed before you run make. Check $(INCDIR) and $(LIBDIR) in particular, and make the choice of ASM or C cursesio version as mentioned above (the distribution version uses the C version of cursesio).
The manual file (curses.man) has been changed at appropriate places.
I would like to thank the following persons for their help:
Brandon S. Allbery (alberry@ncoast.UUCP)
for running comp.binaries.ibm.pc (at that time)
and comp.source.misc.
Steve Balogh (Steve@cit5.cit.oz.AU) for writing a set of manual pages and posting them to the net.
Torbjorn Lindh for finding bugs and suggesting raw character output routines.
Nathan Glasser (nathan@eddie.mit.edu) for finding and reporting bugs.
Ingvar Olafsson (...enea!hafro!ingvar) for finding and reporting bugs.
Eric Rosco (...enea!ipmoea!ericr) for finding and reporting bugs.
Steve Creps (creps@silver.bacs.indiana.edu) for doing a lot of work - among others posting bug fixes to the net, and writing the new cursesio.c module.
N. Dean Pentcheff (dean@violet.berkeley.edu) for finding bugs and rewriting cursesio.asm for Turbo 'C' 1.5.
Finally, Jeff Dean (parcvax,hplabs}!cdp!jeff) (jeff@ads.arpa) has had a shareware version of curses deliverable since about half a year before I released PCcurses 1.0 on Use- Net. He is very concerned about confusion between the two packages, and therefore any references on the network should make clear whether they reference Dean's PCcurses or Larsson's PCcurses.
The changes from v.1.0 to v.1.1 are minor. There are a few bug fixes, and new (non-portable) functions for verbatim IBM character font display have been added (in charadd.c and charins.c). The manual file (curses.man) has been changed at appropriate places.
In the file v10tov11.dif there are listings of the differences between version 1.0 and 1.1. The diff listings are in UNIX diff(1) format.
Version 1.1 compiles with Turbo C v.1.0, as well as Microsoft C v.3.0 and v.4.0. On the release disk there is a make.exe utility which is very similar to UNIX make (If the package was mailed to you, the make utility will be in uuencoded format - in make.uu - and must be uudecoded first). It is much more powerful than Microsoft's different MAKEs; the latter ones will NOT generate libraries properly if used with the PCcurses makefiles.
There are three makefiles:
makefile generic MSC 3.0 makefile makefile.ms MSC 4.0 makefile makefile.tc Turbo C 1.0 makefile
To make a library with for example Turbo C, make directories to hold .H and .LIB files (these directories are the 'standard places'), edit makefile.tc for this, and type
make -f makefile.tc all
and libraries for all memory models will be created in the .LIB directory, while the include files will end up in the .H directory. Also read what is said about installation below!
This is the release notes for the PCcurses v.1.0 cursor/window control package. PCcurses offers the functionality of UNIX curses, plus some extras. Normally it should be possible to port curses-based programs from UNIX curses to PCcurses on the IBM PC without changes. PCcurses is a port/ rewrite of Pavel Curtis' public domain 'ncurses' package. All the code has been re-written - it is not just an edit of ncurses (or UNIX curses). I mention this to clarify any copyright violation claims. The data structures and ideas are very similar to ncurses. As for UNIX curses, I have not even seen any sources for it.
For an introduction to the use of 'curses' and its derivatives, you should read 'Screen Updating and Cursor Movement Optimization: A Library Package' by Kenneth C. R. C. Arnold, which describes the original Berkeley UNIX version of curses. It is available as part of the UNIX manuals. The other source of information is 'The Ncurses Reference Manual' by Pavel Curtis. The latter is part of Curtis' ncurses package.
The only other documentation provided is a 'man' page which describes all the included functions in a very terse way. In the sources, each function is preceded by a rather thorough description of what the function does. I didn't have time to write a nice manual/tutorial - sorry.
PCcurses is released as a number of source files, a man page, and a make file. A uuencoded copy of a 'make' utility, and a manpage for the 'make' is also provided to make it easier to put together PCcurses libraries. Even if you are not interested in PCcurses, it may be worthwhile to grab the make.
The makefile assumes the presence of the Microsoft C compiler (3.0 or 4.0), Microsoft MASM and LIB, plus some MS-DOS utilities. The reason for supplying MAKE.EXE is that the Microsoft 'MAKE:s' are much inferior to a real UNIX make. The supplied make is a port of a public domain make, published on Usenet. It is almost completely compatible with UNIX make. When generating the curses libraries, the makefile will direct make to do some directory creating and file copying, and then re-invoke itself with new targets. The workings of the makefile are not absolutely crystal clear at first sight... just start it and see what it does.
For portability, the curses libraries depend on one assembler file for access to the BIOS routines. There is no support for the EGA, but both CGA, MGA, and the HGA can be used. The libraries are originally for Microsoft C, but all C modules should be portable right away. In the assembler file, segment names probably need to be changed, and possibly the parameter passing scheme. I think Turbo C will work right away - as far as I understand, all its conventions are compatible with Microsoft C.
There are some parts left out between ncurses and PCcurses. One is the support for multiple terminals - not very interesting on a PC anyway. Because we KNOW what terminal we have, there is no need for a termcap or terminfo library. PCcurses also has some things that neither curses nor ncurses have. Compared to the original UNIX curses, PCcurses has lots of extras.
The BIOS routines are used directly, which gives fast screen updates. PCcurses does not do direct writes to screen RAM - in my opinion it is a bit ugly to rely that much on hardware compatibility. Anyone could fix that, of course...
One of the more serious problems with PCcurses is the way in which normal, cbreak, and raw input modes are done. All those details are in the 'charget' module - I do raw I/O via the BIOS, and perform any buffering myself. If an application program uses PCcurses, it should do ALL its I/O via PCcurses calls, otherwise the mix of normal and PCcurses I/O may mess up the display. I think my code is reasonable... comments are welcome, provided you express them nicely...
To install, copy all files to a work directory, edit 'makefile' to define the standard include and library file directory names of your choice (these directories must exist already, and their path names must be relative to the root directory, not to the current one). You must also run uudecode on make.uu, to generate MAKE.EXE. You can do that on your PC, if you have uudecode there, otherwise you can do it under UNIX and do a binary transfer to the PC. When you have MAKE.EXE in your work directory (or in your /bin directory), type make.
Make will now create 4 sub-directories (one for each memory model), copy some assembler include files into them, copy two include files to your include directory, CHDIR to each sub-directory and re-invoke itself with other make targets to compile and assemble all the source files into the appropriate directories. Then the library manager is run to create the library files in your desired library directory. Presto!
If you only want to generate a library for one memory model, type 'make small', 'make large', etc. The name of the memory model must be in lower case, like in the makefile.
I think the package is fairly well debugged - but then again, that's what I always think. It was completed in May-87, and no problems found yet. Now it's your turn... Comments, suggestions and bug reports and fixes (no flames please) to
-- Bjorn Larsson
此处可能存在不合适展示的内容，页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容，可点击提交进行申诉，我们将尽快为您处理。