CVS log for src/lib/libcurses/refresh.c

Up to [cvs.NetBSD.org] / src / lib / libcurses

Request diff between arbitrary revisions

Keyword substitution: kv
Default branch: MAIN

Pull up following revision(s) (requested by rin in ticket #756):

	lib/libcurses/refresh.c: revision 1.127

Don't set the clear_to_eol variable in doupdate to NULL, this was a
leftover from previous code that makes no sense and makes worms(6)
crash randomly.  Thanks to kre@ for reporting this.

Fix typo in comment.

Don't set the clear_to_eol variable in doupdate to NULL, this was a
leftover from previous code that makes no sense and makes worms(6)
crash randomly.  Thanks to kre@ for reporting this.

Revert previous change, it is not correct.

When performing a scroll, check the last line of the region is the same
on virtscr and curscr because the indexes past are supposed to be
one *past* the last matching line (they may actually match if the line is
at the bottom of the screen).  Iff they don't match reduce the scroll
region size by one so we don't scroll non-matching lines, also check
if the region is then 0 after the decrement and just return if it was.

Fixes for lib/56926 amongst other things:
 - plod now correctly accounts for wide characters when plodding
 - use erase line when in color mode if the terminal has the capability
 - ensure that the CA_CONTINUATION flag is applied consistently to the
   subsequent characters in a wide character.
 - fix a bunch of refresh bugs that caused inconsistent placement of
   wide characters.

 * Don't redraw the background if the new background character is the
   same as the old one.  This prevents excessive redraws in some
   applications.

 * Fix bug introduced when wbkgrndset was fixed, we cannot blindly
   replace any instance of the old background character with the new one
   because some of those characters were put there by the application
   leading to display corruption.  So flag characters as background when
   they are erased and only update the flagged characters when setting
   the background.

fix for PR 55496

 * Fix bkgrndset so that it actually sets the background character in
   in line with the SUSv2 specification.

 * Add an internal function to copy a complex character

 * Make the previously static celleq function into a libcurses private
   function so that it can be called in other files.

Fix build (enabled only for MKDEBUG?)

Fix compile breakage when wide char support is disabled.

Make the default colour pair be pair 0 which appears to match other
curses implementations.

Correct (hopefully) the handling of wide characters.

* Remove the WCOL family of macros, these were "stealing" the upper bits
  of a character attribute to store the column width of a character.  No
  warning was given about this in curses.h which meant it was easy to
  accidentally reuse the bits in use by the WCOL macros (we already did).
  Add couple of 16bit ints to the character structure iff HAVE_WCHAR is
  true to hold the display width and wide char related flags (just
  continuation at the moment)
* Convert all instances of WCOL macros to just reference the column width
  in the char structure so it is not obfuscated.
* Fix cursor positioning so placing a cursor in the middle of a wide char
  actually does just that.
* Fix plod so it understands that if the cursor is going to be positioned
  in the middle of a wide char it cannot just reprint the char to get there.
* Fix plodput so it correctly counts the number of output characters for
  wide characters.
* Fix slk routines to properly size the wctomb() buffer.

curses: spell "foreground" correctly in comments

PR lib/56388

For __newwin() and __resizewin(), the line hash was calculated as if
HAVE_WCHAR is disabled.

Fix this bug by refactoring __hash_line() function, which calculates
the line hash by an appropriate method.

Style fixes most for __CTRACE().

Expand __CTRACE() to __nothing #ifndef DEBUG.

Remove most of #ifdef DEBUG around __CTRACE() calls.

No binary changes, except for line numbers for assert().

avoid accessing stack garbage.

on arm64eb resuming vi(1) would often crash.  in makech(), the 'csp'
variable is either set to current window data, or a local stack
variable's address '&blank'.  the window data has many lines of info
stored, and 'csp++' is used per line here.  unfortunately, a case
existed where 'csp++' operated on csp initialised from '&blank' which
eventually crashes when, on my display with 160 columns and 'csp + 155'
exceeds the mapped stack and crashes.

match the '!_cursesi_screen->curwin' conditional that initialises csp,
and avoid csp++ here.  assert() that csp != &blank in both places that
modify csp.

thanks to jdc@ and mlelstv@.

XXX: possibly also should avoid the putch() here as well.

Merge changes from current as of 20200406

0x%p --> %p for non-external codes.

Sync with HEAD

Rework previous fix for getch cursor position when cursor is moved
without refresh.  If the window is not dirty but the window cursor
position does not match curscr then move the cursor.  This fixes
the issues seen in PR lib/54263.

Back out incorrect fix for PR 53617 and fix it in a different way.
Keep track of the cursor location, if getch is called without a refresh
and without pending updates (dirty windows) then move the cursor to the
correct location directly.  Doing this prevents unnecessary refreshes.

Add more debug to _wnoutrefresh

Add/improve debug output for some routines.

Minor adjustment to prior to match putch

Use WCOL rather than wcwidth(3) as we should already know the width
during refresh.
Ensure the character width is not negative when advancing during refresh
(unlikely) and we actually have something to insert in the lower right
corner depending on terminal caps.

Fixes PR lib/54085

Synch with HEAD

_wnoutrefresh - extend current position checks.

Also verify that dwin->cury >= begy (ditto for x), i.e. for pads make
sure that the current position is after the beginning of the displayed
portion.  While here refactor the checks for better readability.

We should probably combine the y and x checks b/c if one of them is
not in the range, the current position as a whole is not in the range
and it doesn't make sense to pick and set just the y or just the x
part of it.

_wnoutrefresh - fix current position checks for pads

maxy is an absolute value and should not be adjusted for begy when we
check also absolute cury.  Note that this change does not affect
calculations for normal windows as for them we always pass zero for
begy.  Ditto for x.

prefresh - fix current position calculations

The code was copy-pasted from wrefresh(), but for pads we need to
adjust for (pbegy, pbegx).

PR lib/53801

Sync with HEAD, resolve a few conflicts

curses: Fix inverted logic when drawing LR corner with auto margins

Thanks to M. Levinson.

Sync with HEAD, resolve a couple of conflicts

Reindent a few lineeq() calls now that we can w/out exceeding 80
columns limit.  Whitespace change only.

Further tweak previous.  Change the check order to be more logical.
No functional change intentded.  While here fix indentation.

Merge one more #ifdef with mostly identical code in both branches.
The same object code is generated modulo gratuitously different order
of checks.

Merge another #ifdefs that now has identical code in both branches
modulo swapped arguments to celleq() call.  Same object code is
generated otherwise for both HAVE_WCHAR and !HAVE_WCHAR

Merge few #ifdefs that now have identical code in both branches.
Same object code is generated for both HAVE_WCHAR and !HAVE_WCHAR.

Define celleq() and lineeq() for !HAVE_WCHAR too and use them instead
of explicit memcmp() calls.  Same object code is generated for both
HAVE_WCHAR and !HAVE_WCHAR, but we can now try to merge some of the
#ifdefs

Rename cellcmp() and linecmp() to celleq() and lineeq() respectively
and make them static.

quickch - fix inverted checks.

I have to admit I have no idea what's going on here, but HAVE_WCHAR
code here inverted the sense of linecmp and cellcmp checks w.r.t the
!HAVE_WCHAR variant just above it.  Cf. also all other similar #ifdefs
in this file.

Found through source code inspection, not really tested.

cellcmp - use logical && not bitwise &

Fix operator precedence for !HAVE_WCHAR

curses: ensure attributes are correctly set for each character

Another fix for PR# 30978.

Sync with head

curses: test terminal capabilities for drawing in the bottom right corner

If auto margins are present, we need to turn them off or enter
an insert mode to "push" the character we want into the corner.
If pushing a wide character of greater than one cell, we need something
more than ich1, othewise the bottom right corner will remain blank.

Follow on fix for PR# 30978.

OK jdc@

Back out the following revision(s) (requested by roy in ticket #1049):

	lib/libcurses/refresh.c: revision 1.89
	lib/libcurses/newwin.c: revision 1.53

  curses: allow drawing the lowest right hand cell of the terminal

  This is a historical behaviour that needs fixing
  If any terminal does scroll when drawing in the lowest right hand cell
  of the terminal then an entry should be made in the terminfo database
  (currently there is no standard code) to state that and define
  __SCROLLWIN as before.

  Fixes PR# 30978.


As discussed in the PR, this change introduces a regression.

Pull up following revision(s) (requested by roy in ticket #1049):

	lib/libcurses/refresh.c: revision 1.89
	lib/libcurses/newwin.c: revision 1.53

curses: allow drawing the lowest right hand cell of the terminal

This is a historical behaviour that needs fixing
If any terminal does scroll when drawing in the lowest right hand cell
of the terminal then an entry should be made in the terminfo database
(currently there is no standard code) to state that and define
__SCROLLWIN as before.

Fixes PR# 30978.

curses: allow drawing the lowest right hand cell of the terminal

This is a historical behaviour that needs fixing
If any terminal does scroll when drawing in the lowest right hand cell
of the terminal then an entry should be made in the terminfo database
(currently there is no standard code) to state that and define
__SCROLLWIN as before.

Fixes PR# 30978.

Sync with HEAD

Sync with HEAD

Make this compile again: use ce instead of clr_eol which gcc thinks might
be NULL.

Sync with HEAD

Kill stray refs to __virtscr and _cursesi_screen missed in prior commit.

ripoffline requires wnoutrefresh to work before _cursesi_screen has been
set.

When doupdate is called, check for typeahead input after N lines changed
instead of aborting really early.

This allows some screen update when holding the page down key for example.

Test for __ISPASTEOL when moving the cursor and it hasn't moved.
Unset __ISPASTEOL when moving the cursor during refresh.

Sync with HEAD.  (Note that most of these changes are simply $NetBSD$
tag issues.)

KNF.
Normalise coding style.
White space police.
Sprinkle some extra braces to make the flow more clear.

No functional changes.

u_int -> unsigned int, u_int32_t -> uint32_t.

Implement POSIX Curses typeahead function.

Clear the "forced" flag after updating a line, otherwise we'll always do
complete line redraws.

Rebase to HEAD as of a few days ago.

sync with head.

for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.

this commit was splitted into small chunks to avoid
a limitation of cvs.  ("Protocol error: too many arguments")

Correct the implementation of mvderwin, it now works as specified by
SUSv2.

Make the touch family of calls actually force an update of the
terminal bypassing optimisations.  Previously if curses thought curscr
was in sync with virtscr (curses concept of what is on the screen) then
nothing would be output.  This change forces an update out to the terminal
regardless.

resync from head

Add extra debugging when copying window contents.

sync with head.

Fix build break if curses is built without wide char support

Use tiparm instead of vtparm.

Catchup with rmind-uvmplock merge.

Correctly test for __ISDIRTY.
Problem noticed by (and test case provided by) blymn@.

Fix an off-by-one error.
Thanks to Matthias Drochner.

Userland now builds and uses terminfo instead of termcap.

OK: core@, jdc@

Prepare curses for the possibility of changing from termcap to terminfo.
term.h #defines lines, pad_char and no_color_video macros which conflict
with existing curses code. We change lines to alines and nlines depending
on use, pad_char to padchar and no_color_video becomes no_color_attributes
but with a strong alias from no_color_video.

Sync w/ -current. 34 merge conflicts to follow.

sync with head.

remove unnecessary casts.

sync with HEAD

Make it build without HAVE_WCHAR

We don't need separate #ifdef's for __ALTCHARSET and WA_ALTCHARSET, as they
are defined to be the same in curses.h.

Fix compile errors on i386 by changing %ld to %zu.

Fix some compiler warnings.
Remove shadowed variable declarations when DEBUG is defined.

Merge in wide curses code done as a Summer of Code project by
Ruibiao Qiu.

Mark subsequent cells of multi-column characters with __WCWIDTH and
their attributes and only display them once.

Output the non-spacing characters to the screen.

Don't overwrite character width when copying characters between screens.

Ansify static function prototypes

 * rearrange code to ensure that SET_WCOL() is called after attributes
   are copied/set.

 * fix some instances where attributes were being checked without
   masking with WA_ATTRIBUTES

 * Replace another instance of non-spacing character copying with function
   call.

Mask off the attribute bits when checking which attributes we need to set.

Only produce debug information for the current line.

* Make WCHAR libcurses the default build
* Sprinkle about some SET_WCOL() calls where they were missing

Apply the changes that add debug "areas" on HEAD to the wcurses branch.
Add correspending changes to the debug code that is only present on the branch.

Add debug "areas" that allow selective debugging by setting the
"CURSES_TRACE_MASK" environment variable.  Postive vales include
debug areas, negative values exclude them.

Wide curses merge

Background characters and attributes don't need to be kept per character
cell, as they are merged when characters are added.  Remove the per cell
storage and clarify the manual page.  Pointed out by ruibiao@.

Don't read entry after end of array.
Discovered by Neil Ludban running curses under 'ElecticFence'

Apply patch (requested by riz in ticket #5516):
Sync libcurses with HEAD of 2005-07-22.

Change first arg of __hash_more to const void * to remove need for double casts

check fflush() return code (from Julian Coleman)

Changes to support *echochar().
Should fix PR lib/24927.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Rework the way window/wub-window refresh is done.

For each line that is marked dirty on each sub-window, we now mark the
relevant line on the parent as dirty and unmark it on the sub-window.
This avoids copying the same data multiple times and also leaves the
cursor in the correct position for the window refreshed.

Make the refresh complete in the presence of multiple sub-windows.
Match function names in the debug output.

Stop refresh(subwin) doing refresh on the other subwin's created before it.
(If nothing else the cursor would get left in the wrong place).
Fix some if (!flags & MASK) checks.

Now refresh() refreshes subwindows too.
Fixes PR lib/20544 by rtr.

Fix resize of windows with subwins.
Let window size be given as a -ver number => lines from bottom.
Fix trace calls of pointers
(approved by Brett Lumn)

Check for conflict between colour and other attributes.  The previous
checks when adding attributes could be inadequate if background colours
were later added to a window.

Back out part of rev 1.50.  We can't call werase() here.
Fixes PR lib/19739 by Matthias Scheler.
Problem also noted by Martin Husemann.

Add pnoutrefresh() and prefresh().
Modify _cursesi_wnoutrefresh() to handle pads.

When copying a line from a window, only copy the changed part instead of
the whole line.
This fixes a bug when displaying displaying overlapping windows.

Fix lint nits mainly with the len parameter type in hash_more.

Set the curses default colours to white on black when using colour.
See the Single UNIX Specification, Version 2 :

  http://www.opengroup.org/onlinepubs/007908799/xcurses/can_change_color.html

Also, add the functions :

	use_default_colors();
	assume_default_colors(fore, back);

(from ncurses) that allow the terminal default colours or user-specified
default colours to be used.

PR/17352: David Laight: Curses core-dumps on windows > 1024 wide.

Remove extraneous tabs from blank lines.

fix __startwin() so it outputs the init strings to the correct device.

* Major change to add support for the newterm/set_term functions.
* Added fix to getch.c suggested by Gabriel Rosenkoetter (thanks :-)

Add code to allow resizing of windows and the underlying terminal.

Rename variables refering to termcap capabilities from NN to __tc_nn.  Case
adjusted to match termcap capability.  A few other variable names renamed too
(ones related to or derived from termcap variables).

pullup 1.39 -> 1.40 (approved by releng-1-5)
on repaint, we must look at attributes present on curscr.
check __ISPASTEOL on curscr->lines, not on win->lines.  PR 10732.

pullup (approved by releng-1-5)

 > use __cputchar(), not putchar().
 > otherwise, __CTRACE() will not be useful for debugging curses internals.

1.18 -> 1.19 basesrc/lib/libcurses/cr_put.c
1.38 -> 1.39 basesrc/lib/libcurses/refresh.c
1.12 -> 1.13 basesrc/lib/libcurses/scroll.c

on repaint, we must look at attributes present on curscr.
check __ISPASTEOL on curscr->lines, not on win->lines.  PR 10732.

use __cputchar(), not putchar().
otherwise, __CTRACE() will not be useful for debugging curses internals.

More include cleanup.  Remvoe (p) from #undef in libipsec.

Sync w/ netbsd-1-5-base.

Move setting of __endwin so that we also change it from the stop handler.

Copy all background attributes (not just colour) from win to __virtscr.

GC __FORCEPAINT.

Fix a silly bug that caused blanks to always be redrawn on dirty lines in
curscr.  This usually manifested itself as a very slow redraw after suspend.

More microoptimization, and explicitly skip lines that are not dirty.

Two optimizations in quickch():
* Don't bother comparing lines that are not dirty when looking for the top
  and bottom regions.
* In the loop that searches for the largest equal region, do the totally
  half-assed hack of splitting the inner loop into two parts -- comparing
  only the hash values the first time, and doing memcmp()s the second time.
  This makes many of my test cases >100x as fast.
  XXX This code needs a lot more work.

We already initialize UP and BC when we fire up Curses, so pass a null pointer
to t_goto() to avoid doing it again... and again... and...

Clear any cached __LEAVEOK flag in __virtscr when refreshing another window.

Don't try and refresh those parts of windows that are outside our terminal.

As the comment suggests, move the __unsetattr() to the end of doupdate(), just
before the fflush().  This avoids needlessly resetting modes.

Speed up refresh in the presence of attributes by precalculating a mask for
each `end' command to turn off other attributes in the case of conflicts.

We can't assume colour won't be turned off if the escape sequences for
turning off attributes aren't the same as the one for colour, so decide
that we've turned off colour when we turn off any attributes.

Remove test code inadvertently left in.

Don't change the cursor position if __LEAVEOK.

Implement wnoutrefresh() and doupdate().  wrefresh() now calls these functions
and the previous wrefresh() code is split between them.  Background character
and attribute handling is now done in wnoutrefresh(), thus simplifying the code
in doupdate(), makech() and quickch().
Refine xterm workround and test for it earlier - this cuts down the number of
lines we test when looking for a scrolled region.
Rename unsetattr() to __unsetattr(), so it can be used by __stopwin().

Work round xterm bug which displays the wrong background colour when the
screen has just been scrolled.
Only use CL if all lines have the same background colour.
Fix debug output in quickch().

Converted all termcap library calls to the "new" interface, this fixes
a problem with curses crashing when the CM capability was larger than
64 bytes and eliminates some possible buffer overflow problems.

Add support for background characters and attributes.
Add extra colour/attributes debugging.

Really don't move with altcharset set.
Leave other attributes if MS is defined.
Move attribute unsetting into a new function.

Don't move with altcharset or colour set (if you try this on an xterm, you
won't actually move, but will think you have).

Added functions to replace what were previously macros in curses.h
(this is a requirement of SUSv2) - the old macro behaviour can be
restored by defining _CURSES_USE_MACROS.
Changed function prototypes to use ANSI style.
All externally visible functions now have ANSI style declarations.

Add color manipulation.
Note that window background attributes are not handled yet.

Made data structures opaque

Fix debug crash when using subwins.

Add alternate character set support.

Changes to make libcurses SUS v2 compatible.

When dumping the attributes for a line, put a space between each attribute
so you can tell where each one starts and stops, and prefix these lines
with "attr: ".

Remove the test for NOQCH, and print the line flags in the CTRACE in
the main loop wrefresh().

- protect NULL attributes from de-referencing
- Silence pointer casts

Upgrades the standard NetBSD curses library to provide some
of the SYSV curses facilities.  The added features are the collapsing
of arrow and function keysequences (as defined by termcap for the
terminal) into symbolic code returns thus relieving the application of
recognising multi-character key sequences.  Other features are the
capability to perform a timed wait for a key (good for when you are
not sure if there is a keypress ready or not) and the capability for
turning off the inter-key timeout when assembling multi-character
function keys.

this work was done by Julian Coleman <J.D.Coleman@newcastle.ac.uk>
and blymn@baea.com.au (Brett Lymn).  i'm just integrating it.  thanks
HEAPS guys!

remove obsolete register declarations

Make endwin() save state so a subsequent call to wrefresh restores curses
and makes endwin()/wrefresh() act like a SIGTSTP.

RCSid police, fix warnings

clean up import

file refresh.c was added on branch netbsd-1-0 on 1994-08-17 21:52:41 +0000

clean up import

new libcurses, from new nvi

repeat after me: "I hate rcs ids"

update to new version from berkeley.  doesn't compile yet, nor
does it have rcsid's.  this is for diffs.

New version from uunet.

Add RCS identifiers.

added Andrew Chernov's patch set:
Standard curses library use eight bit for standout mode, so
8-bit characters displays like highlighted 7-bit characters.

This patch produce library which is fully compatible with all curses
programs and add 8-bit chars to all input/display functions.
---
I don't think, that any programs wish to use internal curses
attribute _STANDOUT directly, in expressions like:
        addch( ch | _STANDOUT );
        Normal interface use standout() and standend() functions instead.
        Many programs use 'char' type (with sign extention) for input characters
        and sign extention becomes _STANDOUT mode in this case.
        So, I refuse this future and allow 8-bit characters for programs,
        which is designed for 7-bit only ('char' type using instead of
        'unsigned char').
---
This small patch fix unpleasant standard curses bug:
curses can't expand TAB at all (but tries).
A man who wrote this curses misplace SYNC_IN and SYNCH_OUT,
this patch exchange macro calls.

This patch useful for standard 7-bit curses too, for this
you must delete '_' symbol before waddbytes and apply patch.
---
Oh, NO! This curses are really buggy!

This small patch fix following problem:
[ assumed scrollok(stdscr, TRUE) ]
when addch(ch) at lower right corner of screen, curses are realy
gone mad instead if simple scrolling... Curses code assumed that
this will be done correctly, but implement it with two bugs.

initial import of 386bsd-0.1 sources

Initial revision