Merge pull request #349 from jordan-hemming/wincon-mousewheel
master
Bill Gray
2025-09-26 10:53:55 -0400
93dba3adf6Fixed apparent typo preventing mouse wheel working with wincon. Changed HOUSE_WHEELED and HOUSE_HWHEELED to MOUSE_WHEELED and MOUSE_HWHEELED.
Jordan Hemming
2025-09-26 11:24:59 +0100
VT/pdcdisp.c modified to compile in non-wide builds (where MAX_UNICODE is not #defined). Removed a no-longer-needed check for dummy characters when drawing text in PDC_transform_line().
Bill-Gray
2025-09-09 14:51:34 -0400
Digital Mars(R) makefile for VT modified, mostly cosmetically, to align with similar makefiles for the DOS and DOSVGA platforms. This also means the 'tests' are now built, and a meaningless dependency on the advapi32.lib library was removed.
Bill-Gray
2025-09-09 14:50:01 -0400
'dosutil.c' misbehaved for negative arguments. Also fixed a timing error of (theoretically) one part in 250 million, and added explanations.
Bill-Gray
2025-09-09 12:27:01 -0400
'naptest.c' revised so that it can be compiled with a wider range of compilers (it previously had some Digital Mars specific bits in it).
Bill-Gray
2025-09-09 12:11:38 -0400
PC_wc_to_utf8() should be declared for export (for use when dynamically linked). 'show_uni' needs to reflect that declaration, and needed a workaround for cases where it's built with 8-bit characters.
Bill-Gray
2025-09-09 11:42:53 -0400
WinGUI didn't consistently reset 'key_modifiers' (state of Ctrl, Alt, Shift) back to zero when a key went up. This meant that Ctrl+key would cause mouse clicks to have Ctrl set (until another Ctrl-less key was hit).
Bill-Gray
2025-09-09 11:06:23 -0400
Framebuffer README.md updated to mention Unifont use for wide-character builds, and some discussion of fullwidth characters.
Bill-Gray
2025-08-31 20:38:50 -0400
'testcurs' color cube and gradient demos shouldn't be used unless we have enough color pairs. Renamed the former demo from 'extended()' to 'show_color_cube()'.
Bill-Gray
2025-08-31 17:06:28 -0400
Through a hack, PSF2 font files can now express fullwidth characters, and the framebuffer/DRM port can display them. See comments in 'hex2psf2.c' in the 'junk' repository for an explanation. (Basically, PSF fonts normally store only one font width; therefore, we now break fullwidth glyphs into 'left' and 'right' halves.)
Bill-Gray
2025-08-31 15:45:21 -0400
Reversion of commit 7d0763d029. I'm at a loss to explain that commit; it appears to undo a perfectly good fix for no good reason?
Bill-Gray
2025-08-31 15:41:33 -0400
The fixes for casting a function pointer used in commits 7309a03c81 and ce60aa6608 don't work on really old Microsoft Visual C (5.0, possibly others)
Bill-Gray
2025-08-10 18:54:27 -0400
Extracted the logic for deciphering combined characters in vt/pdcdisp.c. That same logic may be used in other platforms, and shouldn't be duplicated in each of them.
Bill-Gray
2025-08-09 23:11:19 -0400
'widetest.c' : to determine if combining characters are used (i.e., wchar_t is two bytes), we should test WCHAR_MAX. Instead, I assumed that only Microsoft Windows would have that peculiarity. Issue #340 pointed out (among other things) that at least one MS-DOS compiler (djgpp) also assumes a two-byte wchar_t.
Bill-Gray
2025-08-09 21:19:01 -0400
8eee506d77WinCon/pdcscrn.c: support stdio redirections
Robin Haberkorn
2025-08-09 20:13:25 +0300
VT port in DOS : certain escape sequences (to store/restore the screen, to set the cursor mode, and to set the mouse mode) just produce garbage in DOS. DIVROUND assumes 32-bit arguments; when compiled with 16-bit integers, overflow happened.
Bill-Gray
2025-07-18 21:34:25 -0400
COLORS and COLOR_PAIRS are integers. So if sizeof( int) == 2, we can't have 20 bits for colors. (The only configuration where this appears to be possible is with OpenWATCOM. The Digital Mars compiler supports 16-bit code, but not 64-bit integers.)
Bill-Gray
2025-07-18 21:18:18 -0400
Changed some references to PDCurses to PDCursesMod (situations where the behaviors really are different between the two forks)
Bill-Gray
2025-07-17 14:23:32 -0400
'testcurs' need not call wcslen(). Makes code more sensible and fixes issue #340 (failure to compile with djgpp, which oddly lacks wcslen().)
Bill-Gray
2025-07-17 14:20:19 -0400
Framebuffer/DRM ports do a better job of keeping track of shift/ctrl/alt states for mouse buttons and for key presses. For the latter, we should be able to add NumLock/CapsLock flags using ioctl( STDOUT_FILENO, KDGKBLED, &flags); and checking bits set in 'flags'.
Bill-Gray
2025-07-15 13:27:39 -0400
DRM port assumed the video card would be available as /dev/dri/card0. On my laptop, it's /dev/dri/card1, and still more configurations are possible. It may prove necessary to do a 'proper' search for the default video card (usually assumed to be the one used when booting).
Bill-Gray
2025-07-15 13:23:36 -0400
The IS_*_SURROGATE macros were defined in multiple places; now, they're defined only in curspriv.h. Seems Microsoft has added their own versions of these; we protect against that now (shouldn't actually be a problem, except in the very oddball configuration described in issue #339)
Bill-Gray
2025-07-15 13:18:11 -0400
'event_counts' shadowed a previous instance. Pointed out by Gisle Vanem in issue #339. This triggered a warning in Microsoft(R) Visual C, which should have been an error; added -WX flags. Had it been an error, this wouldn't have slipped by.
Bill-Gray
2025-07-15 13:14:10 -0400
Follow-up to commits 436f1c3e8e, 45e3722c03 : 'a function declaration without a prototype is deprecated in all versions of C [-Wstrict-prototypes]'
Bill-Gray
2025-07-03 22:20:42 -0400
_unlink() isn't defined in OpenWATCOM. Actually, it appears to be (unsurprisingly) basically only a Microsoft® compiler thing.
Bill-Gray
2025-06-28 13:27:40 -0400
Framebuffer/DRM ports : if the mouse is used, Alt/Shift/Ctrl modifiers will be returned. Also fixed a problem with the use of PATH_MAX, a somewhat troublesome, not very portable, and not-always-defined value that we (fortunately) don't really need here anyway.
Bill-Gray
2025-06-28 13:21:23 -0400
Framebuffer/DRM ports : export PDC_SCREEN=n, where n=1, 2, or 3, to cause your programs to start up rotated by 90, 180, or 270 degrees clockwise.
Bill-Gray
2025-06-28 13:19:21 -0400
Follow-up to commit 436f1c3e8e : more clang-18 warnings/errors that 'a function declaration without a prototype is deprecated in all versions of C [-Wstrict-prototypes]'
Bill-Gray
2025-06-26 18:10:28 -0400
Framebuffer/DRM : we can now access either the mouse or the keyboard devices. We'll use the latter in subsequent commits to check for Shift, Alt, Ctrl modifiers.
Bill-Gray
2025-06-25 11:44:41 -0400
Updated README.md for framebuffer/DRM port. Clarified a comment about the 'common' mouse handling code. Updated PDC_VER_DAY and HISTORY.md.
Bill-Gray
2025-06-23 22:31:29 -0400
Fix for clang-18 warning/error : 'a function declaration without a prototype is deprecated in all versions of C [-Wstrict-prototypes]'
Bill-Gray
2025-06-23 18:40:51 -0400
The ncurses interface ignores left and right tilt-mouse events, not left and left (sic) events. Copy/pasta error, I suspect.
Bill-Gray
2025-06-22 21:40:27 -0400
VT (and framebuffer and DRM ports) : call PDC_flushinp() with mouse input in the queue, and you enter an infinite loop!
Bill-Gray
2025-06-22 19:41:25 -0400
'testcurs' modified so that mouseinterval() can be set independently of the mouse event mask. Simplifies testing slightly.
Bill-Gray
2025-06-21 22:49:58 -0400
If an application #defines PDC_NCMOUSE, the behavior of mouse function is fully compatible with ncurses. Wheel up/down is mapped to buttons 4/5 pressed.
Bill-Gray
2025-06-21 22:42:25 -0400
Final steps for new release 4.5.2 : change of version constants in 'curses.h' and 'docs/HISTORY.md', a few recent changes logged in the latter, updated 'docs/USERS.md'.
v4.5.2
Bill-Gray
2025-06-19 16:18:09 -0400
DOS 'PDC_napmsl()' didn't round to the nearest clock tick correctly. The quantity added should have been 23590, about half the divisor 47181. With the incorrect value, naps would be one tick (about 55 milliseconds) shorter than they should be about 40% of the time. Unsurprisingly, nobody noticed. (But it was _wrong_.)
Bill-Gray
2025-06-18 22:18:07 -0400
Revised chtype layout grids to be markdown-friendly, a la wmcbrine/PDCurses@26f39dec9 (except we have a couple more tables that needed reformatting)
Bill-Gray
2025-06-18 19:16:39 -0400
Updated USERS.md to list PDCursesMod-specific attributes and ACS glyphs. Emphasized (in curses.h) that the latter are probably best avoided.
Bill-Gray
2025-06-18 18:54:34 -0400
Simon Sobisch
2025-10-03 12:47:58 -0400