getch() triggered an assert() if a multi-byte character was entered. Fixes issue #335, reported by Robin Haberkorn.
Bill-Gray
2025-06-16 18:24:10 -0400
x11new/Makefile had ~20 lines devoted to DOS, MinGW, and other things that an X11 platform need not think about. Pointed out by Simon Sobisch.
Bill-Gray
2025-06-16 18:11:30 -0400
'ozdemo' : improved ASCII map of Australia, and switched to a more visible color scheme (on some platforms, yellow on green is nearly invisible)
Bill-Gray
2025-06-12 13:41:27 -0400
x11new : hitting Ctrl-Minus or Ctrl-Plus decrements or increments the font size. If the window is maximized, this changes the number of lines/columns (and you need to handle the resize event). Otherwise, the window gets smaller/bigger and the number of lines/columns is retained.
Bill-Gray
2025-06-12 11:11:17 -0400
x11new : some memory can be freed, even if NO_LEAKS isn't defined. free( NULL), it turns out, isn't an error; we need not check for a non-NULL pointer before calling free().
Bill-Gray
2025-05-20 21:24:56 -0400
DRM/framebuffer port's only console code use is to shut off the hardware cursor and turn it back on. We don't really need PDC_puts_to_stdout().
Bill-Gray
2025-05-20 16:52:12 -0400
In the framebuffer/DRM port, outputting codes to set the mouse state accomplishes nothing. (It causes no apparent harm, but it's useless code.)
Bill-Gray
2025-05-20 16:51:34 -0400
I expect to make the 'blink state' a steadily incremented value in at least some ports, instead of being a simple on/off switch. If blinking text occurs over stages, instead of just flashing, it may be less jarring.
Bill-Gray
2025-05-20 16:45:37 -0400
Makefiles for x11new and framebuffer/DRM port were based on the VT makefile. The references to Microsoft Windows NT make no sense here.
Bill-Gray
2025-05-20 16:33:37 -0400
Framebuffer/DRM port can now _show_ (but doesn't yet respond to) a mouse cursor, if built with HAVE_MOUSE=Y. Getting clicks/move events will be a separate, and probably simple, issue.
Bill-Gray
2025-05-20 16:22:45 -0400
Fixed various C90 compliance issues : variables declared after code instead of at the top of a block, true/false used instead of TRUE/FALSE, C++-style comments
Bill-Gray
2025-05-17 20:58:50 -0400
Similar to preceding commit 46a5e0024 : 'false' is not supported by default on OpenWATCOM. (Fixed previously in WinCon, now in WinGUI.)
Bill-Gray
2025-05-08 21:51:32 -0400
VT now uses the 'common/mouse.c' (q.v.) logic for managing mouse input. See preceding commits 52d206870c, 592d27f164, 812bd6953a, and issue #130.
Bill-Gray
2025-05-08 20:45:46 -0400
WinCon now uses the 'common/mouse.c' (q.v.) logic for managing mouse input. See preceding commits 52d206870c, 592d27f164, and issue #130.
Bill-Gray
2025-05-08 20:32:05 -0400
'testcurs' demo revised to display more mouse input data, for both the 'classic' and ncurses interfaces. Added to assist in testing the new mouse-handling code (see commits 52d206870c, 592d27f164, issue #130, and some commits to come).
Bill-Gray
2025-05-08 20:24:22 -0400
WinGUI now uses the 'common/mouse.c' (q.v.) logic for managing mouse input. See preceding commit 52d206870c and issue #130.
Bill-Gray
2025-05-08 20:12:16 -0400
Added code for the platform-independent bits of mouse handling, and revised 'x11new' to use it. Part of issue #330.
Bill-Gray
2025-05-08 20:03:04 -0400
'mouse.c' : if we're trapping the PDCurses*-specific BUTTONn_MOVED events, these are reflected in the ncurses-style 'bstate' bitmask as 'REPORT_MOUSE_POSITION'.
Bill-Gray
2025-05-08 16:51:27 -0400
Moved a couple more screen-related externals (_default_foreground_idx, _default_background_idx) into the SCREEN structure, where they properly belong. Follow-up to preceding commit a60b63df74.
Bill-Gray
2025-04-27 10:50:19 -0400
PDC_blink_state should be (and now is) part of the SCREEN structure. (Many externals should be part of SCREEN. This is a start.)
Bill-Gray
2025-04-27 10:33:50 -0400
The function to mark cells of particular attributes for refreshing really should be static to 'color.c'. (At least at present. There is the possibility that bold or blinking text may need to be similarly marked somewhere, sometime, on some platform(s).
Bill-Gray
2025-04-27 10:27:10 -0400
WinCon/WinGUI PDC_clearclipboard() function now returns PDC_CLIP_SUCCESS only if the EmptyClipboard() call succeeds, aligning with wmcbrine/PDCurses@26f091f0b6 and wmcbrine/PDCurses@da691a62eb . Also some minor reformatting of associated documentation, again to bring it into alignment with William McBrine's version.
Bill-Gray
2025-04-27 10:26:28 -0400
Made PDC_realloc_array() non-static. There are various places within PDCursesMod where it could, or should, be used.
Bill-Gray
2025-04-17 17:50:53 -0400
'testcurs' can now test the ncurses mouse interface. In ncurses, that's all it _can_ test (because ncurses lacks the 'classic' SysV undocumented mouse functions). In PDCursesMod, you can choose at compile time.
Bill-Gray
2025-04-16 17:52:39 -0400
VT port now works (mostly) on Haiku. The Haiku terminal sends different sequences for cursor keys. The sequences sent for mouse movements/presses/clicks are slightly non-standard, but in a way that can still be parsed correctly with slightly changed code.
Bill-Gray
2025-04-14 13:03:26 -0400
WinGUI : if mouseinterval( 0); has been called, and/or click events are not enabled for a particular mouse button, then return mouse button presses immediately; don't queue them and wait for a corresponding button release. From issue #330.
Bill-Gray
2025-04-14 12:40:32 -0400
'testcurs' modified to allow full control over the mouse mask, and to be able to set mouseinterval(0). Largely for testing issue #330 (and related problems noticed while investigating that issue).
Bill-Gray
2025-04-14 12:23:33 -0400
Changes for ncurses test program building : those requiring the forms library (cardfile, demo_forms, dup_fields, form_driver_w) don't work in wide mode and are, at least for the nonce, excluded.
Bill-Gray
2025-04-13 13:02:15 -0400
x11new : the (global) font structure can be static to 'pdcscrn.c'. Preparatory to some improvements in font handling for this platform.
x11new
Bill-Gray
2025-04-13 12:33:55 -0400
x11new : Clarified README.md about use of printf()s in PDCursesMod (and PDCurses) platforms with an underlying terminal
Bill-Gray
2025-04-13 12:32:35 -0400
x11new : Ctrl-C now breaks the program when not in raw-input mode, and FUNCTION_KEY_ABORT works as well if set (tested with 'brk_test' and 'ripoff', respectively)
Bill-Gray
2025-04-03 22:11:07 -0400
x11new and framebuffer/DRM : when blinking, you could get packets longer than MAX_PACKET_LEN. They need to be broken up... which is easy to do.
Bill-Gray
2025-03-28 15:21:31 -0400
'ripoff' demo provides an error message/explanation if you ask for too many lines, instead of just bombing out. Program keeps going, albeit without the requested lines, in both PDCursesMod and ncurses.
Bill-Gray
2025-03-26 21:50:30 -0400
x11_new : need to XSync() in PDC_doupdate(). Screen updates done in PDC_transform_line() may be, and usually are, deferred (see comments in IMPLEMNT.md).
Bill-Gray
2025-03-26 21:08:51 -0400
'New' X11 port : found/worked around remaining X11 memory leaks. Updated docs for this port, and incremented version month/day.
Bill-Gray
2025-03-26 12:15:34 -0400
Font metrics determined via XQueryFont(), rather than being hardwired. This should simplify changing fonts. Also, we ought to (and now do( shift fonts vertically by the font descent.
Bill-Gray
2025-03-22 16:25:01 -0400
Restarting Curses failed to reset the 'prev_fg'/'prev_bg' values. You could get odd colors in the 'restart' test program.
Bill-Gray
2025-03-20 21:19:36 -0400
Framebuffer/DRM port : cursor did not rotate when the screen did. Now, it rotates and we have at least some other cursor styles (half-block, cross, caret sort of) supported.
Bill-Gray
2025-03-12 22:14:10 -0400
DRM port : can specify screen to use via PDC_SCREEN as an integer, and can (with some permissions required) cycle displays by hitting Alt-=
Bill-Gray
2025-03-11 16:49:03 -0400
Array bound checking should be done _before_ checking for null terminators. Similar issue as in commit c03e650a7020c8c2945d74fb4ecf25; in both cases, Valgrind will complain about the access of an element just past the end of the valid array.
Bill-Gray
2025-02-11 21:11:03 -0500
Moved 'minor new features' above 'bug fixes' (which matches the order for previous HISTORY.md entries). Eliminated a redundant paragraph about the removal of PDC_set_box_type().
Bill-Gray
2024-10-28 10:38:57 -0400
WinGUI display : I don't think the code that displays a preceding and trailing character is really needed to avoid stray pixel garbage. In fact, I can't see how that could possibly have ever worked. It caused some (not all) of the flickering problem described in issue #322.
Bill-Gray
2024-10-03 10:23:28 -0400
The 'ripoff' test program now also tests PDC_set_function_key(). There wasn't previously a demo or test that exercised that function.
Bill-Gray
2024-10-03 10:19:42 -0400
WinCon and WinGUI Borland Turbo C makefiles now specify the lack of a 'stdint.h'. If you build with -DUTF8, you need not add -DWIDE (this has been the case with most makefiles for a long time, but we'd not caught up on the Borland ones until now).
Bill-Gray
2024-07-29 20:14:57 -0400
The ncurses 'menu' library can now be built and used with PDCursesMod, and the ncurses program 'demo_menus' built and used to test it out.
Bill-Gray
2024-07-29 12:49:34 -0400
Simon Sobisch
2025-06-18 12:16:07 -0400