971a7f9c14I like seeing my name and all, but I don't think these comments are needed anymore.
William McBrine
2005-12-15 17:32:00 +0000
efef844f7a"supports" -> "claims to support". Some people are confused when they see this message, but also see that blinking doesn't happen.
William McBrine
2005-12-15 17:08:03 +0000
0a41b05838Redundant setting of direct_video to TRUE just before checking the environment variable negated the check in PDC_query_adapter_type(), which sets direct_video to FALSE on CGA to avoid snow. See the mailing list archive for more. To Do: Needless to say, not only is BIOS mode slow to begin with, but CGA cards are typically found only in the slowest machines, so PDCurses' performance there is terrible. Snowless direct video for CGA is possible, but difficult.
William McBrine
2005-12-15 00:56:49 +0000
d936eb3c56Always update cursor in PDC_gotoxy(). For reasons unclear, characters would sometimes be skipped in BIOS mode if the update wasn't done, even when it appeared unneeded. See the mailing list archive for more.
William McBrine
2005-12-15 00:48:15 +0000
7e26778580Missing files in the tar command line made "make dist" fail. Corresponding entries also removed for zip, tar.
William McBrine
2005-12-14 23:53:18 +0000
590c75a8e7Added slk_color.
William McBrine
2005-12-14 21:41:02 +0000
8141cffaaeMore unneeded memmove references.
William McBrine
2005-12-14 21:37:49 +0000
ce392c2cc0Useless memmove() ifdef -- memmove() is not used here.
William McBrine
2005-12-14 21:06:15 +0000
0c8271d467Guard against duplicate declarations in term.h and here. This is another thing that only showed up on my Mac, and I don't get why -- term.h is the same on both.
William McBrine
2005-12-14 19:57:50 +0000
1da495103fI am a dope. Pre-ANSI C doesn't have "const".
William McBrine
2005-12-14 19:40:29 +0000
3af787115fconst-ified panel prototypes (and a couple char * -> void *).
William McBrine
2005-12-13 06:01:22 +0000
96792ce954Corrected "clean" target for Cygwin.
William McBrine
2005-12-13 05:59:08 +0000
f42097b130Added color_set and wcolor_set.
William McBrine
2005-12-13 04:32:30 +0000
1d7693fa23The implicit wrefresh() call in wgetch() was not applied if nodelay was set. This differed from other implementations -- you can see the effect in the "rain" demo. I assume it was meant as a performance boost, but I don't think it makes a significant difference.
William McBrine
2005-12-13 04:27:42 +0000
68e20cb5d2Restored the ability to build the ncurses test programs, using the latest ncurses source.
William McBrine
2005-12-12 23:32:22 +0000
adc25d46beAdded slk_color to curses*.def.
William McBrine
2005-12-12 14:02:59 +0000
28fbae1c92Added [w]color_set to curses*.def.
William McBrine
2005-12-12 13:59:14 +0000
3216ec8e89Added comments about color_set().
William McBrine
2005-12-12 06:17:59 +0000
cbf2ccc15eMacros for mvhline, mvwhline, mvvline, and mvwvline, as found in the standard. To Do: Functions?
William McBrine
2005-12-12 04:22:32 +0000
f44348acedslk_color(), after slk_attrset(). Not tested.
William McBrine
2005-12-12 04:09:49 +0000
361618047bAdded wcolor_set() and color_set().
William McBrine
2005-12-12 00:26:58 +0000
5563422512Added ptest.exe for Cygwin, removed unused lines.
William McBrine
2005-12-11 23:04:56 +0000
995079a138Cygwin claims to HAVE_USLEEP, but it doesn't work. I took out the define so it falls back to the Win32 call, which does. Also, Cygwin wants HAVE_STRING_H.
William McBrine
2005-12-11 23:03:04 +0000
2007481023"Line-breakout optimization" now defaults to off (i.e., typeahead(-1)). It has never really worked correctly, AFAICT, and it's one of the leading causes of complaints about PDCurses. I've made the minimum necessary change -- just setting the flag to FALSE on startup. So, if someone actually wants this, they can still enable it with typeahead(0).
William McBrine
2005-12-11 05:51:24 +0000
63c79216d3XSetInputFocus() causes BadMatch errors in Gnome, and just in general, seems to do more harm than good.
William McBrine
2005-12-11 02:03:49 +0000
180959e77eReintegrated the duplicative PDC_breakout()s, preparatory to fixing it. To Do: Is this the best place for the reunified function?
William McBrine
2005-12-11 01:07:19 +0000
1497a57fa5Code that depended on the value of ERR and OK -- ugh! I was alerted to this by a suggested patch in the ECurses (Eiffel Curses) package -- but the patch actually makes it worse, with the current ERR/OK values. Instead, I changed the affected private functions to use booleans.
William McBrine
2005-12-10 21:44:45 +0000
aea421806bAdded PDC_DLL_BUILD, commented out PDC_THREAD_BUILD. The DLL works now in LCC with this change.
William McBrine
2005-12-10 15:29:12 +0000
1686bdf1adReplaced PDC_STATIC_BUILD with its opposite, PDC_DLL_BUILD (so the default is NOT to build a DLL). Suggestion of Ron House. Quite right, too. Also, commented out "-MT -DPDC_THREAD_BUILD". The DLL build actually works in MSVC, without these options; but I'm still leaving it commented out for now.
William McBrine
2005-12-10 14:24:59 +0000
0160f804a3Fix for bug #657543 on SourceForge -- cursor not turning off in full-screen mode. SetConsoleCursorInfo() gets confused by a height of 1 (see analysis by "nobody"); so instead, we'll just toggle the visibility flag without attempting to set the height, which works fine.
William McBrine
2005-12-09 16:38:36 +0000
ade543e9e9Upgraded version numbers.
William McBrine
2005-12-09 01:54:50 +0000
e245fa03ecRemoved obsolete comment about the non-implementation of immedok().
William McBrine
2005-12-09 00:07:31 +0000
1bd1de34a6BouncingBalls() was needlessly re-init_pair()ing and wattrset()ting on every pass through the loop.
William McBrine
2005-12-08 17:27:08 +0000
a601ca9cdawin->_maxx etc. is not standard, and is defined differently in PDCurses vs. ncurses (off by one), so use getmaxyx() instead. Sadly, getmaxy() isn't standard either, which means undoing a warning cleanup.
William McBrine
2005-12-08 16:33:39 +0000
32c11502d3subpad() was erroneously offsetting from the origin coordinates of the parent pad (which are always -1,-1), causing an off-by-one error. testcurs modified accordingly to look the same -- I figure the "now" shouldn't be covered. (The pad test now looks the same in ncurses, etc.)
William McBrine
2005-12-08 08:12:32 +0000
0104d50a07Some more neatening.
William McBrine
2005-12-08 06:12:11 +0000
e66e860b4aNeatened up a little.
William McBrine
2005-12-08 05:48:04 +0000
a8343ad111PDC_gattr() does not exist. PDC_gattr() has never existed. PDC_gattr() will never exist.
William McBrine
2005-12-07 11:56:05 +0000
45e6236661As long as we're changing ERR to -1 (seems unnecessary to me, but anyway), we really ought to make OK be 0, too, as in other curses implementations (ncurses, BSD, Sun).
William McBrine
2005-12-07 11:53:19 +0000
f13417cd68In wborder(), whline(), and wvline(), when dealing with characters without their own defined colors, the current (wattrset) attribute would be added to them, but not the current background (wbkgd). This was visible in the testcurs demo. The same issue was fixed in PDC_chadd() in 1993. This also eliminates some redundant code.
William McBrine
2005-12-07 09:52:06 +0000
0830405b5dRemoved wclrtoeol() from Continue() -- it seemed to serve no purpose except to break the box on introTest().
William McBrine
2005-12-07 06:12:48 +0000
613689f36eOops... ifdef -> ifndef XCURSES for resizeTest proto.
William McBrine
2005-12-07 05:25:35 +0000
91f0d52d73Standardized the prototyping for the demos, plus some minor changes to ensure they all compile under ncurses. (Everything works there, except some of the borders.)
William McBrine
2005-12-07 01:37:43 +0000
05b8b3c45fThe behavior of the testcurs tests varied depending on the order in which they were launched. I've added a few werase() and box() calls to rectify this, along with an extra pause in the scroll test so that the first message can be more easily read. But I'm less than confident in these changes... I need to better determine the original intent.
William McBrine
2005-12-06 23:36:33 +0000
3fd8d1f3c7Fixed the menu-drawing optimization so it actually works. Also took out A_PROTECT, because I think it's hideous.
William McBrine
2005-12-06 10:29:38 +0000
a7f7040288A_REVERSE was not working correctly in XCurses. This is actually two fixes: First, due to flipped parameters in the assignment in the second section where A_REVERSE was checked in XCursesDisplayText(), in some cases, A_REVERSE didn't work at all. You could see this in the opening menu of the testcurs demo if you moved the highlight bar. Second, even in the "correct" section, A_REVERSE was hardwired to black-on-white, instead of actually reversing the current color pair. This was fixed for the other ports some time ago, and now for XCurses. To Do: Fix the remaining redundant code in XCursesDisplayText(). (It took me longer to find this than it should have, because of that.)
William McBrine
2005-12-06 10:25:51 +0000
5014ba4836A few more signed/unsigned comparisons. More to come.
William McBrine
2005-12-06 01:56:18 +0000
b0e742f3b9Minor warning cleanup (comparison of signed and unsigned).
William McBrine
2005-12-06 01:23:31 +0000
107431eb9bXCursesProgramName is no longer used; added Xinitscr() calls in its place, where possible (not in tuidemo).
William McBrine
2005-12-06 00:20:31 +0000
fc1021235fRedundant declaration of trap(); added check of sig as warning supression.
William McBrine
2005-12-05 23:32:15 +0000
fd253bac06Warning cleanup for 64-bit systems (no difference for 32) -- cast call_data (orignally XtPointer, aka void *) as long instead of int.
William McBrine
2005-12-05 23:17:37 +0000
7c53d533f4Changed color pairs 8, 9 and 10 in BouncingBalls() to 11, 12 and 13. 8-10 were already used in SubWinTest(), which isn't a problem in itself; but, prior to this change, if PDC_PRESERVE_SCREEN was set, the screen was still cleared when the exisiting color pairs were reinitialized. With this change, only the central box is touched.
William McBrine
2005-12-03 05:52:35 +0000
25e15e4c3eMore warning cleanup; turned on all warnings for LCC (now that there are no more). init_pair() could use more reworking, but my more ambitious revision was buggy, so this will do for now.
William McBrine
2005-12-03 04:49:32 +0000
c7edc0268aWarning cleanup -- last of the ANSI prototypes. (To Do: Do we really still need to support non-ANSI compilers?)
William McBrine
2005-12-02 23:47:52 +0000
230ff73841Missing underscore in initTest() prototype section.
William McBrine
2005-12-02 22:21:55 +0000
35a94b8274Use static linking for the demos in XCurses, so they can be tested without installing. (Dynamic link options retained as comments.)
William McBrine
2005-12-02 01:16:22 +0000
02b6a51df3Successfully tested the 16-bit chtype with several 32-bit compilers, so I removed the comment about that. I could've sworn it was broken before, but I guess it was something else I'd changed in the scratch copy.
William McBrine
2005-12-01 19:49:22 +0000
81dc1d2b61On 64-bit systems, the X11 port was ignoring every other keystroke. This is because XCursesSendKeyToCurses() was mistakenly pushing an int instead of a long. (If there are any 16-bit X11 systems out there, they'd also be affected.)
William McBrine
2005-12-01 16:44:46 +0000
ee818de2e5Minor clarification in PDCurses Text Attributes comments.
William McBrine
2005-11-30 18:19:51 +0000
a392bf38f4Revised the "PDCurses Text Attributes" comment section for consistency and completeness.
William McBrine
2005-11-30 18:13:38 +0000
95dbcefa14DLLs don't work right in MSVC, either, so disable them for now. Also, STARTUP and "-I." are unneeded.
William McBrine
2005-11-29 20:03:44 +0000
7ad41570f6CC_HOME etc. not needed here, plus it interfered with compilation under the gcc3.2.1 upgrade package for EMX.
William McBrine
2005-11-29 19:02:05 +0000
2f35f7fb46Added ptest, improved clean target, commented out lxlite call.
William McBrine
2005-11-29 18:37:41 +0000
a0e719e872CCINCDIR not needed here.
William McBrine
2005-11-29 17:55:11 +0000
a719e38df8Another ptest.exe build added.
William McBrine
2005-11-29 17:42:00 +0000
f1a59fb0efAdded some empty definitions of A_INVIS, etc., for 16-bit mode (i.e., with CHTYPE_LONG not defined). This is needed to get it to compile. (That done, it still works, with 16-bit compilers -- tested with Borland 3.1 and Open Watcom 1.3).
William McBrine
2005-11-29 08:35:30 +0000
7a65ea9b5cAdd "/out:" for panel.lib; add "-ansic" to avoid any inadvertent LCCisms.
William McBrine
2005-11-27 17:28:16 +0000
63b5ffa2bdRemoved unused vars. Also, might as well disable the DLL as well as its demos, until it's fixed.
William McBrine
2005-11-27 07:36:25 +0000
50b261f7e4Added ptest.exe build for LCC.
William McBrine
2005-11-26 06:29:27 +0000
96929c87ecRemoved redundant declaration of done().
William McBrine
2005-11-24 19:04:41 +0000
cc384bbba5LCC has INT_MAX in limits.h, and doesn't want it in curses.h. ("ifndef" doesn't work, since limits.h is included after curses.h.)
William McBrine
2005-11-24 19:02:01 +0000
f2fa74a1b4In PDC_vsscanf(), n sometimes holds negative values, so it should not be unsigned. (Warning cleanup for LCC.)
William McBrine
2005-11-24 18:56:57 +0000
90607f50d8Improved "clean" target.
William McBrine
2005-11-24 18:54:38 +0000
93fef41cc5Warning cleanup: prototypes for malloc, etc.
William McBrine
2005-11-24 18:50:40 +0000
ccfff63b1cAdded a few more PDCURSES_SRCDIRs, and took out a few unneeded CC_HOMEs. More of these to do.
William McBrine
2005-11-24 00:56:41 +0000
138496105chalfdelay() to the rescue again.
William McBrine
2005-11-23 23:46:42 +0000
4deb6dde48PDC_BUILD is no longer PDC_BUILD????, so reflect that in the comments. To Do: It's probably about time to update that build number.
William McBrine
2005-11-23 19:44:20 +0000
febc7c19a2Duplicate GetLastError() calls.
William McBrine
2005-11-23 15:30:52 +0000
dd3d18b984One more thing to delete in "clean": *.res.
William McBrine
2005-11-23 14:49:52 +0000
234b9d4f58Fix for bug 1144353
Mark Hessling
2005-11-23 04:42:43 +0000
2f457e6c60Move idle() to front of waitforkey() so we don't have to wait for the clock; also, comments.
William McBrine
2005-11-23 01:16:55 +0000
78dfd0d86dtuidemo was incredibly slow in XCurses (due to a "sleep(1)" after each keystroke!), and pegged the CPU on other platforms. The correct solution is halfdelay(). It's also much simpler.
William McBrine
2005-11-23 00:47:07 +0000
7c354194f0Got rid of meaningless casts, eliminated a warning.
William McBrine
2005-11-22 15:44:02 +0000
49e575bfcaIS_THIS_REALLY_NEEDED? No. Minor warning cleanups.
William McBrine
2005-11-22 05:55:04 +0000
cba1f2318aSuppressed all the remaining Watcom warnings, and most of the ones in Borland, via command-line options. The Watcom warnings were all about the cbreak()-style macros -- "Expression is only useful for its side effects." Nothing to be done about that one. Borland had a similar one ("Code has no effect"), along with a lot of warnings about unused parameters, where functions haven't been fleshed out yet (and in some cases, probably never will be). The latter may be useful, but we probably don't need to see them every time.
William McBrine
2005-11-22 04:49:07 +0000
23843d8d74Another ptest build added.
William McBrine
2005-11-22 01:04:08 +0000
b25c067ff7Build ptest along with other demos. More to come; I only added it for the ones I've successfully tested.
William McBrine
2005-11-22 01:03:24 +0000
d9d24acb4fWarning cleanup (mainly added prototypes).
William McBrine
2005-11-22 00:45:45 +0000
54465b4700Removed path from CCFLAGS -- not needed, plus the line became too long for tuidemo to actually be built under DOS. Now it works.
William McBrine
2005-11-21 23:37:43 +0000
91e0b153bbImproved "clean" targets. More to do in this regard.
William McBrine
2005-11-21 20:49:57 +0000
edce1d7c3eBorland's make already knows what CC is. By omitting it, the file works for both Turbo and Borland compilers without further editing.
William McBrine
2005-11-21 19:46:37 +0000
c804ec55cfRedundant, unused files.
William McBrine
2005-11-21 19:40:52 +0000
e6abbfed15New format for time logging -- the inclusion of CLOCKS_PER_SEC was adding a dependency on math libraries for old Borland DOS compilers, which (among other things) broke the demo builds.
William McBrine
2005-11-21 19:39:03 +0000
8dfdc53fd1Grammar fix: "required" -> "require".
William McBrine
2005-11-21 04:36:14 +0000
5f72dc3c00Small correction: int -> attr_t.
William McBrine
2005-11-21 04:03:03 +0000
9788129736Color-changing code was really messed up. See bug #1227795 on SourceForge for more.
William McBrine
2005-11-21 02:59:58 +0000
4b364d193eMore warning cleanup.
William McBrine
2005-11-21 01:03:37 +0000
f56ebed424Added some error-checking.
William McBrine
2005-11-20 18:55:11 +0000
59037b58a7Changed 500 to 680 in color_content() to match ncurses. To Do: This should report actual values for XCurses (and init_color() should work there).
William McBrine
2005-11-20 18:44:16 +0000
William McBrine
2005-12-15 17:32:00 +0000